The DOI matches its Crossref record; 219 of the 219 references it cites were found in a registry or an index; 3 of the 77 sources its claims cite were checked for support.
Explainable, trustworthy, and ethical machine learning for healthcare: A survey
Open accessPublished version of a preprint
Abstract
With the advent of machine learning (ML) and deep learning (DL) empowered applications for critical applications like healthcare, the questions about liability, trust, and interpretability of their outputs are raising. The black-box nature of various DL models is a roadblock to clinical utilization. Therefore, to gain the trust of clinicians and patients, we need to provide explanations about the decisions of models. With the promise of enhancing the trust and transparency of black-box models, researchers are in the phase of maturing the field of eXplainable ML (XML). In this paper, we provided a comprehensive review of explainable and interpretable ML techniques for various healthcare applications. Along with highlighting security, safety, and robustness challenges that hinder the trustworthiness of ML, we also discussed the ethical issues arising because of the use of ML/DL for healthcare. We also describe how explainable and trustworthy ML can resolve all these ethical problems. Finally, we elaborate on the limitations of existing approaches and highlight various open research problems that require further development.
Overview
AI snapshot
Written from the paper’s full text.
TL;DR
This survey reviews explainable, trustworthy, and ethical machine learning for healthcare, covering XML methods, security/robustness challenges, and bioethical issues. It proposes a development-to-deployment pipeline and argues explainability is needed to build clinician/patient trust and address ethical/legal requirements.
Key findings
- The survey distinguishes explainability from interpretability, taxonomizes XML into intrinsic/post-hoc, model-specific/agnostic, surrogate, and visualization methods, and reviews healthcare applications of DT, rule-lists, RF, GAM, PDP, CAM/GradCAM, LRP, LIME, DeepLIFT, SHAP, SA, GBP, IG, AM, HSCNN, P2V, and EMANet.
- It proposes a pipeline from development to deployment: data explication, data annotation, feature extraction, black-box or inherently explainable model, XML method, validation, clinical validation, and clinical implementation.
- It identifies trustworthiness barriers including adversarial attacks, privacy breaches, distribution shifts, data imperfections, and a trade-off among accuracy, explainability, and robustness; it cites a study of 18 ImageNet classifiers showing an accuracy–robustness trade-off.
- Ethical review covers principlism (beneficence, non-maleficence, autonomy, justice; explicability added for AI) and concerns such as imbalanced data, data bias, privacy, informed consent, and care ethics; examples include the Google Health breast cancer study accused of transparency/reproducibility violations and racial bias in healthcare spending data.
Methods
- Design: narrative survey/review of explainable, trustworthy, and ethical ML for healthcare, with comparison tables of prior surveys and XML methods.
- Data/sample: published literature and cited healthcare ML/XML studies; no primary dataset or new empirical experiment.
- Analysis: thematic synthesis of XML techniques, security/safety/robustness issues, ethical frameworks, pitfalls, and future research directions.
Limitations
- Paper-stated limitations of existing approaches: no formal definitions, lack of standardized representation and requirements, inadequate validation metrics, lack of theoretical understanding and causality, ethical constraints, and security challenges; pitfalls include explanations vulnerable to input changes, sub-optimal explanations, dependence on data/model, unclear accountability, and need for rigorous evaluation.
- Evident: as a survey, it does not conduct a systematic search protocol, meta-analysis, or primary validation; the ethics review is explicitly selective because “Due to space availability, we will not be able to provide a comprehensive review of all the relevant publications.”
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
69 Not checked yet · 50 No basis given · 41 Own interpretation · 3 Couldn’t confirm · accuracy not yet measured
The cited sources that could be checked couldn’t confirm or dispute this claim.
In the paper
“They showed that SHAP explanations to class probability did not improve the decision-making.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > B. Sub-optimal Explanations
What the sources say
- Could not verifycites [171] Lee, 2012, Ethical justification for conducting public health surveillance without patient consent, Am J Public Health sourceExperimental — accuracy not yet measured
The Conclusions discuss the ethical justification for public health surveillance without patient consent. Searches of our copy for SHAP explanations, class probability, and alert-processing decision-making found no passage addressing the claimed result. (LLM, conf=0.78)
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Could not verify
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“To the best of our knowledge, there is no review that covers these topics in-depth while highlighting the link and applications of explainable and trustworthy methods for the medical domain.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > I. INTRODUCTION
hedged- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: unclear — It is unclear why the paper says it holds.
In the paper
“In addition, this black-box nature of the DL is yet a major challenge in developing trustworthiness [25].”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > G. Lack of Theoretical Understanding
What the sources say
- Not checked yetcites [25] Gille, 2020, What we talk about when we talk about trust: Theory of trust for AI in healthcare, Intell.-Based Med.
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“The current field of medical XML lacks requirement guidelines about designing, measuring, and testing explanations.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > C. Lack of Standardized Requirements for XML
What the sources say
- Not checked yetcites [22] Wierzynski, 2018, The challenges and opportunities of explainable AI, Intel. Com
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“Also, there is no measure to check which explanations should be preferred from the different explanations produced by the same method [24].”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > F. Validation of Explanations
What the sources say
- Not checked yetcites [24] Gilpin, 2018
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“many recent studies have highlighted the vulnerabilities of these systems towards adversarial ML attacks [30].”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > J. Security Challenges
What the sources say
- Not checked yetcites [30] Szegedy, 2013
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“However, due to the unavailability of sound fundamental laws and models, the explanation of DL cannot be produced up-to the required scale.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > G. Lack of Theoretical Understanding
What the sources say
- Not checked yetcites [8] Begoli, 2019, The need for uncertainty quantification in machine-assisted medical decision making, Nat. Mach. Intell. source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“entrusting the decisions of black-box systems that are not explainable can be life-threatening [9]”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > I. INTRODUCTION
What the sources say
- Not checked yetcites [9] Holzinger, 2017
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“Similar is the case for data leakage, which can mislead the model learning and testing [49].”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > Validation Method:
What the sources say
- Not checked yetcites [49] Robnik-Šikonja, 2018, Perturbation-based explanations of prediction models
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“ML and XML algorithms are designed to aid the medical staff, not for replacing the medical experts [45].”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > C. Need of XML for Healthcare
What the sources say
- Not checked yetcites [45] Kaufman, 2012, Leakage in data mining: Formulation, detection, and avoidance, ACM Trans. Knowl. Discov. Data source
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Probably not a substantial claim. 99
- It may state two separate claims as one. 36
The cited sources that could be checked couldn’t confirm or dispute this claim.
In the paper
“They also stated that these factors were not revealed using traditional techniques.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > d) Local Interpretable Model-Agnostic Explanations (LIME):
What the sources say
- Could not verifycites [113] Palatnik de Sousa, 2019, Local interpretable model-agnostic explanations for classification of lymph node metastases, Sensors sourceExperimental — accuracy not yet measured
Sections 3.2 and 3.5 describe tissue characteristics highlighted by LIME explanations, but they do not say those characteristics were unrevealed by traditional techniques. A search for comparisons with traditional or conventional techniques found no passage establishing that claim. The citing sentence also does not identify what “these factors” refers to. (LLM, conf=0.72)
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
- Could not verify
The cited sources that could be checked couldn’t confirm or dispute this claim.
In the paper
“The implementation of P2V improves the predictive model working efficiency and also increase the explainability of these models.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > l) Patient2Vec (P2V):
What the sources say
- Could not verifycites [122] H. Jung, Y. Oh, Towards better explanations of class activation mapping, in: Proceedings of the IEEE/CVF International Conference on Computer Vision, 2021, pp. 1336–1344. sourceExperimental — accuracy not yet measured
The Conclusion (section 5) reports improved visual explanations and a single backward pass for LIFT-CAM, not P2V. Searches for P2V, predictive-model efficiency, and explainability found no passage in our copy addressing the claimed effects of P2V. (LLM, conf=0.76)
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
- Could not verify
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“Despite the increasing research on the said topic, few researchers focused on the problem of evaluating XML.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > 4) Measuring the effectiveness of explanations:
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.
In the paper
“However, these models have black-box nature and lack underlying theoretical foundation behind their decisions [8]”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > I. INTRODUCTION
What the sources say
- Not checked yetcites [8] Begoli, 2019, The need for uncertainty quantification in machine-assisted medical decision making, Nat. Mach. Intell. source
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
In the paper
“This human-machine combination is a powerful tool to facilitate the patients and develop high-quality treatments [21].”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > C. Need of XML for Healthcare
What the sources say
- Not checked yetcites [21] M.A. Ahmad, C. Eckert, A. Teredesai, Interpretable machine learning in healthcare, in: Proceedings of the 2018 ACM International Conference on Bioinformatics, Computational Biology, and Health Informatics, 2018, pp. 559–560. source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: fails — It may be too vague to check against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. It may be too vague to check against evidence. The paper’s conclusions may not depend on it. It may state two separate claims as one.
In the paper
“The explainable techniques allow the researchers or end-users to get involved in improving the performance of the algorithm and to trust the prediction results [44].”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > C. Need of XML for Healthcare
What the sources say
- Not checked yetcites [44] Caruana, 2015, Intelligible models for HealthCare: Predicting pneumonia risk and hospital 30-day readmission
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
None of the sources cited for this claim have been checked yet.
In the paper
“These insights can help improve the prescribed medicines, facilities provided to patients in hospitals [46], and health monitoring systems [47].”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > C. Need of XML for Healthcare
What the sources say
- Not checked yetcites [46] Guidotti, 2018, A survey of methods for explaining black box models, ACM Comput. Surv. source
- Not checked yetcites [47] Doshi-Velez, 2017
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Due to the non-monolithic concept and subjective nature of explainability, evaluation of explanations is a complicated task.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > 4) Measuring the effectiveness of explanations:
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. It may state two separate claims as one.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“This approach does not require humans in the loop (layperson or domain expert).It works appropriately when human-based or application-based evaluations have already been performed.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > b) Human-based Evaluation:
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds. It may state two separate claims as one.
In the paper
“The blackbox methodology provides local explanations of a model for a prediction [41].”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > B. Taxonomy of XML
What the sources say
- Not checked yetcites [41] Chen, 2021, Ethical machine learning in healthcare, Annu. Rev. Biomed. Data Sci. source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“However, the difference is that it does not require a costly experimental environment and domain expert for testing.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > b) Human-based Evaluation:
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds. It may state two separate claims as one.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“An XML technique usually explains in a human-understandable way how the feature of data relates to prediction results, i.e., what features of X-ray images a model learns to detect the fractures.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > E. Characteristics of XML for Healthcare
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“It is now evident that the explainability of black-box models is required to attain fair and trustworthy healthcare decisions.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > D. Enhancing the Clinical Practice of ML : A Framework of Effective XML for Healthcare
boosteddirect- Asserted: passes — The authors put it forward as true.
- Checkable: fails — It may be too vague to check against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. It may be too vague to check against evidence. The paper may give no reason why it holds.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“Low translucency has the advantage of more compact explanation results.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > Characteristics of XML model for healthcare
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“This technique helps in evaluating the explanation in real-time practical scenarios.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > 4) Measuring the effectiveness of explanations:
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
In the paper
“These properties are mostly required for explanations of any black-box model, however, we are describing these in terms of healthcare domain.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > E. Characteristics of XML for Healthcare
What the sources say
- Not checked yetcites [51] Işın, 2016, Review of MRI-based brain tumor image segmentation using deep learning methods, Procedia Comput. Sci. source
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
In the paper
“Robnik et al. [51] listed some properties of good quality XML method.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > E. Characteristics of XML for Healthcare
What the sources say
- Not checked yetcites [51] Işın, 2016, Review of MRI-based brain tumor image segmentation using deep learning methods, Procedia Comput. Sci. source
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
In the paper
“The model-based explanation methods are well developed and implemented for healthcare applications (further discussed in Sec IV-D).”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > Validation Method:
What the sources say
- Not checked yetcites [50] Rasheed, 2020, Machine learning for predicting epileptic seizures using EEG signals: A review, IEEE Rev. Biomed. Eng.
boosted- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“For end-users of the medical field, i.e., clinicians, radiologists, pathologists, neurologists, and patients, it is more likely that they do not have enough knowledge of understanding complex mathematical explanations.So, for them, rulebased, textual, or visualization-based explanations are required.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > Characteristics of XML model for healthcare
hedgeddirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds. It may state two separate claims as one.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“One can test the explanations with laypersons, and it helps to generalize the findings as the more number of testers (laypersons) are easily available.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > b) Human-based Evaluation:
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“High translucency allows the explanation method to gather information from more internal parameters of a model.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > Characteristics of XML model for healthcare
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
In the paper
“Thus, the explanation provides a fair analysis of model architecture and learned parameters [42].”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > C. Need of XML for Healthcare
What the sources say
- Not checked yetcites [42] Levin, 2021, Machine-learning-based hospital discharge predictions can support multidisciplinary rounds and decrease hospital length-of-stay, BMJ Innov. source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: fails — It may be too vague to check against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. It may be too vague to check against evidence. The paper’s conclusions may not depend on it.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“It is how an explainable model represents its explanation according to the application domain and end-users.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > Characteristics of XML model for healthcare
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“It represents how much the explaining method depends on the internal architecture of the ML model, i.e., learnable parameters.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > Characteristics of XML model for healthcare
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
- Rests on an assumption: There is no definite method to calculate the properties of good quality XML methods.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“The signal method is used to highlight the patterns of data that activate the neurons of higher layers.This can be done by back-propagating a signal from the last layer of the network to the input layer.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > F. Explaining DL Techniques
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“To get a generalized model for clinical use low translucency is desirable and for a patient-specific application, the explanation should be more detailed and accurate that can lead towards high translucency.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > Characteristics of XML model for healthcare
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“It illustrates the variety of ML methods for which an explanation technique can be applied.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > Characteristics of XML model for healthcare
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“A good explanation should be accurate specially in case of disease prediction.Low value of accuracy is acceptable if the performance of ML model is also low.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > Characteristics of XML model for healthcare
plaincorrelational- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds. It may state two separate claims as one.
In the paper
“This local explanation allows the end-user to determine which feature is important for the precise prediction and how the perturbation in feature affects the prediction results.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > F. Explaining DL Techniques
What the sources say
- Not checked yetcites [64] Samek, 2021, Explaining deep neural networks and beyond: A review of methods and applications, Proc. IEEE source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“These techniques are precise in terms of performance but their decisions are difficult to explain because of their complex model architecture.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > F. Explaining DL Techniques
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. It may state two separate claims as one.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“Saliency methods and feature attribution (FA) are the two broad categories of methods applied for explanation of DL models.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > F. Explaining DL Techniques
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“Fidelity is a property of explanation that shows the precision of approximating the decision of ML method.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > Characteristics of XML model for healthcare
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Explanation methods of complex deep neural networks (DNN) have high translucency thus can not be applicable for other models.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > Characteristics of XML model for healthcare
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“Saliency methods produce the explanation by presenting important feature maps of each data sample.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > F. Explaining DL Techniques
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
In the paper
“They showed that the heatmaps produced by SA are much more noisy compared to the heatmaps generated using LRP [65].”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > F. Explaining DL Techniques
What the sources say
- Not checked yetcites [65] Bohlin, 2006
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“Gradient-based saliency methods reveal how the output of the model changes with a small change in the input.These methods are computationally efficient because of a single pass of input (forward and backward) through the network.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > F. Explaining DL Techniques
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“It is crucial for a typical ML/DL empowered healthcare system to be fully trusted by both the clinicians and patients for getting the real impact of such systems.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > F. Explaining DL Techniques
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: fails — It may be too vague to check against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. It may be too vague to check against evidence. The paper may give no reason why it holds.
In the paper
“LIME produces an explanation list that shows the contribution of individual features to the prediction.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > F. Explaining DL Techniques
What the sources say
- Not checked yetcites [64] Samek, 2021, Explaining deep neural networks and beyond: A review of methods and applications, Proc. IEEE source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
None of the sources cited for this claim have been checked yet.
In the paper
“The literature refers to two sets of popular principles that have been outlined by the Organisation for Economic Co-operation and Development (OECD) [67] and European Commission's AI High-Level Expert Group (HLEG) [68] that can be used for sustaining trust and trustworthiness in AI.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > A. Principles of Trustworthy AI for Healthcare
What the sources say
- Not checked yetcites [67] Holzinger, 2022, Information fusion as an integrative cross-cutting enabler to achieve robust, explainable, and trustworthy medical artificial intelligence, Inf. Fusion source
- Not checked yetcites [68] Mirsky, 2019, CT-GAN: Malicious tampering of 3D medical imagery using deep learning
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“The techniques with low translucency are applicable to a wide range of ML methods.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > Characteristics of XML model for healthcare
plaincorrelational- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“The feature attribution method decomposes each value produced from each neuron of the output layer according to the contributions made by the individual dimensions of an input sample.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > F. Explaining DL Techniques
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
In the paper
“They concluded that a high concentration of cfDNA in plasma and CNV in chromosomes affected the pathogenesis of cancer cases.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > d) Local Interpretable Model-Agnostic Explanations (LIME):
What the sources say
- Not checked yetcites [118] Bengio, 2010, Decision trees do not generalize to new variations, Comput. Intell. source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.
In the paper
“Using SHAP values they described that the intraoperative urine output, IV fluid infusion, blood product transfusion, and dynamic changes of hemodynamic features are significant causes of injury.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > d) Local Interpretable Model-Agnostic Explanations (LIME):
What the sources say
- Not checked yetcites [113] Palatnik de Sousa, 2019, Local interpretable model-agnostic explanations for classification of lymph node metastases, Sensors source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
In the paper
“They analyzed the decision of CNN using LIME and described that CNN learned the patterns of peripheral lesion of DN glomeruli for DN detection.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > d) Local Interpretable Model-Agnostic Explanations (LIME):
What the sources say
- Not checked yetcites [112] Jo, 2020, Deep learning detection of informative features in tau PET for Alzheimer’s disease classification, BMC Bioinformatics
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“The adversarial robustness can be defined as the survivability of ML-based systems to adversarial attacks.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > B. Secure, Safe, and Robust ML for Healthcare
plain- Asserted: passes — The authors put it forward as true.
- Checkable: fails — It may be too vague to check against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. It may be too vague to check against evidence. The paper’s conclusions may not depend on it.
In the paper
“They concluded that the DLIME shows better results than LIME.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > d) Local Interpretable Model-Agnostic Explanations (LIME):
What the sources say
- Not checked yetcites [116] Pianpanit, 2019
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
None of the sources cited for this claim have been checked yet.
In the paper
“It is worth noting that the majority of the guidelines in both aforementioned frameworks mainly focus on the AI aspects of robustness, safety, security, explainability, and fairness and are therefore the key requirements for building trustworthy AI systems.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > A. Principles of Trustworthy AI for Healthcare
What the sources say
- Not checked yetcites [67] Holzinger, 2022, Information fusion as an integrative cross-cutting enabler to achieve robust, explainable, and trustworthy medical artificial intelligence, Inf. Fusion source
- Not checked yetcites [68] Mirsky, 2019, CT-GAN: Malicious tampering of 3D medical imagery using deep learning
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“The safe and robust ML is a broad term and we define the robustness of the ML/DL models along three dimensions, i.e., robustness to security threats, robustness to distribution shifts, and data imperfections.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > B. Secure, Safe, and Robust ML for Healthcare
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
In the paper
“Zafar et al. [116] pointed out the problem of instability of generated explanations due to perturbation addition and random feature selection in the medical computer-aided diagnosis (CAD) systems.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > d) Local Interpretable Model-Agnostic Explanations (LIME):
What the sources say
- Not checked yetcites [116] Pianpanit, 2019
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“From the figure, it is evident that the trustworthy ML can only be possible by addressing challenges related to privacy, fairness, explainability, security, and robustness.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > B. Secure, Safe, and Robust ML for Healthcare
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim.
In the paper
“The literature suggests that ML systems are not safe, secure, and robust.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > B. Secure, Safe, and Robust ML for Healthcare
What the sources say
- Not checked yetcites [32] Qayyum, 2021, Secure and robust machine learning for healthcare: A survey, IEEE Rev. Biomed. Eng. source
hedged- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“This algorithm works on two axioms sensitivity and implementation variance.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > d) Local Interpretable Model-Agnostic Explanations (LIME):
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“To resolve the problem of toxic drugs a chemist need the structural element which is causing the problem.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > d) Local Interpretable Model-Agnostic Explanations (LIME):
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“IG does not require any changes to the primordial deep neural network.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > d) Local Interpretable Model-Agnostic Explanations (LIME):
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.
In the paper
“They also compared the performance of their proposed model with CNN and concluded that the HSCNN out-performed the CNN with a high prediction score and explainable results.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > d) Local Interpretable Model-Agnostic Explanations (LIME):
What the sources say
- Not checked yetcites [121] Guisan, 2002, Generalized linear and generalized additive models in studies of species distributions: setting the scene, Ecol. Model. source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“IG can be implemented for any kind of model, i.e., image.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > d) Local Interpretable Model-Agnostic Explanations (LIME):
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“In the drug development classification of toxic and non-toxic drugs is not enough.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > d) Local Interpretable Model-Agnostic Explanations (LIME):
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“If the model outcome has changed notably with perturbations, it shows us that the feature has a high contribution to the prediction.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > d) Local Interpretable Model-Agnostic Explanations (LIME):
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.
In the paper
“They showed that the network is sensitive to intensity and sphericity in coherence with domain information.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > d) Local Interpretable Model-Agnostic Explanations (LIME):
What the sources say
- Not checked yetcites [119] Mohnen, 2020, Healthcare expenditure prediction with neighbourhood variables–A random forest model, Stat. Politics Policy source
plaincorrelational- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
In the paper
“For example, the health insurance and portability and accountability (HIPAA) act assures the privacy of the medical data of patients.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > C. Main Ethical Questions and Challenges
What the sources say
- Not checked yetcites [157] Panesar, 2019
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“However, ML algorithms requires access to as much data as possible to improve the precision and accuracy of the outcome.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > C. Main Ethical Questions and Challenges
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“Another problem is that ML-based healthcare products are manufactured by Western companies and these products are developed and tested on Caucasian data.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > C. Main Ethical Questions and Challenges
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds. It may state two separate claims as one.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Explanation of the data, as we proposed in the pipeline of explainable ML presented in Figure 3, in the first place is required to check for these biases.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > C. Main Ethical Questions and Challenges
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Against this background, the quality of training data has a high impact on the performance of ML algorithms.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > C. Main Ethical Questions and Challenges
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. The paper may give no reason why it holds.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“Imbalance class data is one of the common data-related problems that occur in the supervised training of ML/DL models.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > C. Main Ethical Questions and Challenges
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.
In the paper
“This bias in the availability of facilities is also reflected in the data and can cause the biased decision of ML algorithms [155].”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > C. Main Ethical Questions and Challenges
What the sources say
- Not checked yetcites [155] Bohr, 2020
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“It is a common practice that more healthcare facilities are available for wealthy people which makes it less likely that low-income people are able to access advanced technological treatments.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > C. Main Ethical Questions and Challenges
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“it can improve care quality, cut the overall costs, reduce or even eliminate diagnostic errors and improve the process of predicting disease.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > V. ETHICAL ML FOR HEALTHCARE
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds. It may state two separate claims as one.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“ML models learn the latent variable of data to deduce the predictions.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > C. Main Ethical Questions and Challenges
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“Protecting the privacy of patients and the confidentiality of their data is one of the fundamentals of ethical healthcare.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > C. Main Ethical Questions and Challenges
plain- Asserted: passes — The authors put it forward as true.
- Checkable: fails — It may be too vague to check against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. It may be too vague to check against evidence. The paper’s conclusions may not depend on it.
In the paper
“HIPAA's policy standards are designed to improve the healthcare systems and mandate it for all healthcare organizations to protect medical information [157].”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > C. Main Ethical Questions and Challenges
What the sources say
- Not checked yetcites [157] Panesar, 2019
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“The integration of AI/ML into healthcare practice and clinical applications promises to provide substantial improvements to the healthcare sector.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > V. ETHICAL ML FOR HEALTHCARE
plainconditional- Asserted: passes — The authors put it forward as true.
- Checkable: fails — It may be too vague to check against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. It may be too vague to check against evidence. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“This problem arises due to the nonuniform distribution of samples among classes.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > C. Main Ethical Questions and Challenges
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Biases in outcomes of models used for healthcare services may have profound consequences.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > C. Main Ethical Questions and Challenges
hedgedconditional- Asserted: passes — The authors put it forward as true.
- Checkable: fails — It may be too vague to check against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. It may be too vague to check against evidence. The paper may give no reason why it holds.
In the paper
“It is to be noted here that explicability is a newly proposed principle, within the particular AI context, which has the same meaning outlined above in this paper [127].”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > V. ETHICAL ML FOR HEALTHCARE
What the sources say
- Not checked yetcites [127] Preuer, 2019, Interpretable deep learning in drug discovery
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
None of the sources cited for this claim have been checked yet.
In the paper
“To achieve the potential of”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > C. Main Ethical Questions and Challenges
What the sources say
- Not checked yetcites [144] Association, 2001, World Medical Association Declaration of Helsinki. Ethical principles for medical research involving human subjects, Bull. World Health Organ.
- Not checked yetcites [146] Berner, 2007
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“The amount and type of the needed data are increasing by time to the extent of seriously blurring the boundaries between what is”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > C. Main Ethical Questions and Challenges
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: fails — It may be too vague to check against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. It may be too vague to check against evidence. The paper’s conclusions may not depend on it.
In the paper
“The proposed architecture not only able to detect the diseases but also show transparency by highlighting the affected area detected by the system.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > l) Patient2Vec (P2V):
What the sources say
- Not checked yetcites [123] Zafar, 2021, Deterministic local interpretable model-agnostic explanations for stable explainability, Mach. Learn. Knowl. Extract. source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.
In the paper
“In addition, bias in the data is also a major challenge to handle and that can eventually lead to algorithmic bias [178].”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > A. Explaining Medical Data
What the sources say
- Not checked yetcites [178] Wang, 2019, Bias also matters: Bias attribution for deep neural network explanation
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“To make the situation more complex, conducting proper analysis for all these data would necessitate giving access not only to one's physician but to many other experts in various areas.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > C. Main Ethical Questions and Challenges
hedgeddirect- Asserted: passes — The authors put it forward as true.
- Checkable: fails — It may be too vague to check against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. It may be too vague to check against evidence. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Various factors can put people's privacy at risk, and we highlight here two of them: a) Unprotected Data Sharing:”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > C. Main Ethical Questions and Challenges
plainconditional- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
In the paper
“Moreover, it has been argued in the literature that traditional statistical performance metrics like the area under the curve may not be sufficient for evaluating the models making clinical decisions [176].”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > E. Rigorously Evaluating the Method
What the sources say
- Not checked yetcites [176] Anwar, 2018, Medical image analysis using convolutional neural networks: a review, J. Med. Syst. source
hedged- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“ML techniques build their decisions on the latent variables which are learned from the data.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > A. Explaining Medical Data
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. The paper may give no reason why it holds.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“Techniques based on Internetof-Things (IoT) are widely used in healthcare systems for real-time monitoring of critical patients.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > C. Main Ethical Questions and Challenges
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“With the advanced technologies, the records and reports of patients are converted into electronic health records (EHR).”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > C. Main Ethical Questions and Challenges
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“However, this ability leads to data breaching through tracking and monitoring of patient's routines which dishonors the patients' privacy.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > C. Main Ethical Questions and Challenges
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“Medical data is one of the most difficult data to handle due to its complex, multi-variate, and sometimes non-stationary and scarce nature.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > A. Explaining Medical Data
plain- Asserted: passes — The authors put it forward as true.
- Checkable: fails — It may be too vague to check against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. It may be too vague to check against evidence. The paper may give no reason why it holds.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“The dependence of latent-variables on each other can cause misleading patterns and due to this issue, the ML-based decision-making will be misleading.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > A. Explaining Medical Data
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: unclear — It is unclear why the paper says it holds.
It may state two separate claims as one.
In the paper
“In addition, the inherent bias in the input (medical) data (e.g., class imbalance) can be reflected in the model's outputs i.e., model might prefer a specific class as compared to other classes and this bias might influence the explanations of the model [170].”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > A. Vulnerability to Input Changes
What the sources say
- Not checked yetcites [170] Klingler, 2017, Ethical issues in public health surveillance: a systematic qualitative review, BMC Public Health source
hedgedconditional- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“These records are available online via the cloud servers.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > C. Main Ethical Questions and Challenges
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“As these explanations are dependent on the choice of reference point and a slight change in the reference point of gradient will significantly change the explanation thus causing confusion that will eventually lead to misleading results or interpretation.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > C. Dependence on Data and Model
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
In the paper
“This blame game will eventually foster”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > D. Accountability Attribution
What the sources say
- Not checked yetcites [175] Weizenbaum, 1976
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“However, despite their significant performance these models are black-box and lack theoretical understanding behind their decisions.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > VI. POTENTIAL PITFALLS
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
In the paper
“These biases are hard to undo and their elimination have unintended consequences on the results [179].”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > A. Explaining Medical Data
What the sources say
- Not checked yetcites [179] Weerts, 2019
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“It is well established that the explanation of the ML/DL techniques is required to gain the trust of clinicians in ML/DL empowered healthcare solutions.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > B. Representation Techniques for Explanation
boosteddirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. The paper may give no reason why it holds.
- Rests on an assumption: Explanation of ML/DL techniques is required to gain the trust of clinicians in ML/DL empowered healthcare solutions.
In the paper
“In this regard, a few studies suggested involving all types of stakeholders into the ML/DL method development process that may include clinicians, policymakers, data scientists, ML researchers, and hospital staff, to name a few [182], [176].”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > E. Interdisciplinary Development Workforce
What the sources say
- Not checked yetcites [182], [176]
hedged- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“In the literature, visualization-based methods are widely applied to explain the decisions of ML/DL methods.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > B. Sub-optimal Explanations
plaincorrelational- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.
In the paper
“Even if this automation of caregiving will not result in causing medical harm to the patient or job cuts in the healthcare staff, replacing human care will still have social costs, e.g., exchanging feelings and emotions among humans will cease to be part of caregiving [166].”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > C. Main Ethical Questions and Challenges
What the sources say
- Not checked yetcites [166] Atlam, 2019
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: fails — It may be too vague to check against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. It may be too vague to check against evidence. The paper’s conclusions may not depend on it.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“As we discussed in Section III, that the explanations of the DL models are model-specific in nature, therefore, it is also required to develop inherently explainable and generalize explainable methods for the DL algorithms in future.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > C. Generalized Explanations
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Their black-box nature makes them susceptible to many vulnerabilities such as adversarial attacks, biased decisions, and being not able to generalize out of distribution samples, etc.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > VI. POTENTIAL PITFALLS
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: unclear — It is unclear why the paper says it holds.
It may state two separate claims as one.
In the paper
“Over the past few years, it has been shown that ML/DL methods can be easily fooled and desired outcomes can be obtained [30], [31].”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > D. Adversarially Robust ML
What the sources say
- Not checked yetcites [30], [31]
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Moreover, the clinical impact of ML/DL advancements is only completely possible by overcoming challenges like adversarial ML attacks.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > D. Adversarially Robust ML
boosteddirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“However, the developed explanation methods have unique challenges and limitations associated with them, which are described below.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > VI. POTENTIAL PITFALLS
plaincorrelational- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. The paper may give no reason why it holds.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“The critical nature of healthcare applications provides significant motivation for the malicious actors to defame the ML/DL-based system and to get the desired outcomes.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > D. Adversarially Robust ML
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: fails — It may be too vague to check against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. It may be too vague to check against evidence. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.
In the paper
“In the literature, a wide variety of adversarial ML attacks have been already proposed and the research on developing respective defense methods is very limited [73].”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > D. Adversarially Robust ML
What the sources say
- Not checked yetcites [73] Takabi, 2016, Privacy preserving multi-party machine learning with homomorphic encryption
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Such an interdisciplinary development workforce will enable collaboration between the knowledge experts (i.e., clinicians and ML researchers) and healthcare service providers that will eventually improve productivity and outcomes.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > E. Interdisciplinary Development Workforce
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: fails — It may be too vague to check against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. It may be too vague to check against evidence. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.
In the paper
“Obtaining an informed consent from the patient before exposing him/her to any medical intervention is one of the practical applications of these principles [135].”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > C. Main Ethical Questions and Challenges
What the sources say
- Not checked yetcites [135] Miles, 2005
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“As mentioned above, modern bioethics have other approaches besides principlism.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > C. Main Ethical Questions and Challenges
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.
In the paper
“Because of the particular significance of addressing the data-related concerns, different countries have developed policies and laws [160].”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > C. Main Ethical Questions and Challenges
What the sources say
- Not checked yetcites [160] Haibe-Kains, 2020, Transparency and reproducibility in artificial intelligence, Nature source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“There is no doubt that the deployment of ML for clinical practice will aid the clinicians.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > D. Accountability Attribution
boosteddirect- Asserted: passes — The authors put it forward as true.
- Checkable: fails — It may be too vague to check against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. It may be too vague to check against evidence. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“The issues that can be discussed within this approach go beyond the question of solely measuring the efficacy and safety of certain applications or calculating their possible healthrelated benefits and harms.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > C. Main Ethical Questions and Challenges
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“This situation becomes even more complex when we consider all stakeholders in the loop.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > D. Accountability Attribution
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“The advancements in ML/DL techniques have a great potential to revolutionize healthcare.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > E. Interdisciplinary Development Workforce
boostedconditional- Asserted: passes — The authors put it forward as true.
- Checkable: fails — It may be too vague to check against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. It may be too vague to check against evidence. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.
- Rests on an assumption: Advancements in ML/DL techniques have a great potential to revolutionize healthcare.
In the paper
“Lack of ethical guidelines and fewer or no proper training of such surveillance programs raises ethical concerns [162].”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > C. Main Ethical Questions and Challenges
What the sources say
- Not checked yetcites [162] West, 2020
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“As outlined above, the respect for persons and autonomy are among the widely-agreed upon principles in modern bioethics.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > C. Main Ethical Questions and Challenges
boosted- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.
In the paper
“The care ethics approach, which focuses on the domain of intimate human relationships rather than the abstract application of rules [135], serves as a good example in this regard.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > C. Main Ethical Questions and Challenges
What the sources say
- Not checked yetcites [135] Miles, 2005
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“Due to the availability of implantable devices, it is now possible to monitor the patients without their consent.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > C. Main Ethical Questions and Challenges
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“For example, humancomputer interaction (HCI) is a well-developed technique to empower users.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > B. Representation Techniques for Explanation
boosted- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.
- Rests on an assumption: HCI is a well-developed technique to empower users.
The paper relies on this without support.
DL models produce more accurate results than conventional ML techniques.
“These complex and dense DL models produce more accurate results than conventional ML techniques.”
The paper relies on this without support.
Simple ML is easy to explain with less accurate results and complex DL produce more accurate results with fewer explanations due to their complex non-linear structures.
“It is a well-known problem of XML that the simple ML is easy to explain with less accurate results and complex DL produce more accurate results with fewer explanations due to their complex non-linear structures.”
The paper relies on this without support.
There is no consistency in the use of terms like feature relevance, feature importance, saliency maps, and heatmaps.
“There is also the need for defining terms like feature relevance, feature importance, saliency maps, heatmaps, etc.As there is no consistency in the use of these terms.”
The paper relies on this without support.
It is not standardized whether radiologists or neurologists are interested in visualization-based explanations.
“However, it is not standardized yet whether the radiologists or neurologists are interested in these explanations or not.”
The paper relies on this without support.
The measures to validate the quality of produced explanations are not adequate.
“The measures to validate the quality of produced explanations are not adequate.”
The paper relies on this without support.
There does not exist any standard method for measuring the quality of explained healthcare decisions.
“There does not exist any standard method for measuring the quality of explained healthcare decisions.”
The paper relies on this without support.
DL is not efficient to infer causal relations between decisions and data.
“However, DL is not efficient to infer causal relations between decisions and data.”
The paper relies on this without support.
Causal links are required for taking fair decisions.
“These causal links are required for taking fair decisions.”
The paper relies on this without support.
The excellent performance of ML/DL cannot be evidence of its safety.
“As the excellent performance of a ML/DL cannot be evidence of its safety, which is simply the determination of how safe is the ML/DL empowered system for humans, i.e., patients.”
The paper relies on this without support.
The employment of ML algorithms for clinical practice or use of patients is still distant.
“However, the employment of these algorithms for clinical practice or the use of patients is still distant.”
The paper relies on this without support.
The primary reason for the gap between healthcare ML research and clinical practice is the unexplained results of algorithms and sometimes poor performance.
“The primary reason for this gap is the unexplained results of algorithms and sometimes the poor performance of the algorithm.”
The paper relies on this without support.
Any bias in data, subjectivity, redundancy, or problem in data representation causes misleading results.
“Any bias in the data, subjectivity, redundancy, or problem in data representation causes misleading results.”
The paper relies on this without support.
If ML learns from data of patients denied medical care because of no health insurance, it will generate biased results.
“Another example can be the data of patients who were denied to take medical care because of no health insurance.If ML learns from that data, it will generate biased results.”
The paper relies on this without support.
If understandable explanations are given, a doctor is far better prepared to make a decision using these explainable models.
“It is clear that if understandable explanations are given, a doctor is far better prepared to make a decision using these explainable models.”
The paper relies on this without support.
Humans typically have foreknowledge of the problem domain, which they will use to believe or deny a prediction if they understand the explanation of results by the algorithm.
“Humans typically have foreknowledge of the problem domain, which they will use to believe or deny a prediction if they understand the explanation of results by the algorithm.”
The paper relies on this without support.
There is no definite method to calculate the properties of good quality XML methods.
“The only limitation is that there is no definite method to calculate these properties.”
The paper relies on this without support.
Healthcare has unique legal, regulatory, and ethical challenges that must be considered when integrating ML/DL into clinical settings while ensuring deployed systems are robust and free from algorithmic bias.
“Moreover, unlike other domains, healthcare has unique challenges that involve legal, regulatory, and ethical challenges that need to be considered while integrating ML/DL based algorithms into actual clinical settings while ensuring that the deployed systems are robust and free from algorithmic bias.”
The paper relies on this without support.
Data imperfections in real-world data result in biased training and increase false positives and negatives.
“These data imperfections will eventually result in biased training and will increase the false positives and negatives.”
The paper relies on this without support.
The explainability of black-box models can assure reliable and ethical use in the medical field.
“The explainability of black-box models can assure reliable and ethical use in the medical field.”
The paper relies on this without support.
Class activation maps are only applicable for a specific CNN architecture that has a dense and global averaging pooling layer after the last convolutional layer.
“However, CAMs are only applicable for specific CNN architecture, i.e., CNN must have a dense and global averaging pooling layer after the last convolutional layer.”
The paper relies on this without support.
Online healthcare websites are not always concerned about the privacy of users' data.
“These websites also provide free data storing facility and are not always concerned about the privacy of the users' data.”
The paper relies on this without support.
Explanation of ML/DL techniques is required to gain the trust of clinicians in ML/DL empowered healthcare solutions.
“It is well established that the explanation of the ML/DL techniques is required to gain the trust of clinicians in ML/DL empowered healthcare solutions.”
The paper relies on this without support.
HCI is a well-developed technique to empower users.
“For example, humancomputer interaction (HCI) is a well-developed technique to empower users.”
The paper relies on this without support.
Advancements in ML/DL techniques have a great potential to revolutionize healthcare.
“The advancements in ML/DL techniques have a great potential to revolutionize healthcare.”
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
In the citing paper
“The application of AI in healthcare brings forth several limitations that need to be addressed for its successful integration into clinical practice [37].”
From Ethical Framework for Harnessing the Power of AI in Healthcare and Beyond (2024) · cites this paper as [37]
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
Each contradiction between two passages listed here was confirmed by an independent check.
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > d) Local Interpretable Model-Agnostic Explanations (LIME):
“DeepLIFT 2020 Yang et al. [100] Uses a reference value and measures the reference values of all neurons using a forward and backward pass -Genetic variants caused by diseases Single-Nucleotide Polymorphisms”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > d) Local Interpretable Model-Agnostic Explanations (LIME):
“Sharma et al. [117] implemented DeepLIFT for Genome-Wide Association Studies (GWAS), which focused on studying genetic variants caused by common diseases.”
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > D. Applications of XML Models for Trustworthy Healthcare
“these models are also implemented for predicting the depression of Alzheimer's patients [91]”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > References
“91. H. Byeon, "Developing a random forest classifier for predicting the depression and managing the health of caregivers supporting patients with alzheimer's disease,"”
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > B. Key Works
“In Ethics, computing and medicine: Informatics and the transformation of health care, published in 2007 [138]”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > References
“138. K. W. Goodman, Ethics, computing, and medicine: informatics and the transformation of health care. Cambridge University Press, 1998.”
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > d) Local Interpretable Model-Agnostic Explanations (LIME):
“TABLE IV: Summary of model-agnostic explanation methods for healthcare applications.”
§ Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey > D. Applications of XML Models for Trustworthy Healthcare
“However, CAMs are only applicable for specific CNN architecture, i.e., CNN must have a dense and global averaging pooling layer after the last convolutional layer.”
Citations
Cites
Grad-CAM: Visual Explanations from Deep Networks via Gradient-Based Localization
Selvaraju, R. R., Cogswell, M., Das, A., Vedantam, R., Parikh, D., & Batra, D. (2017) · 2017 IEEE International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2017.74 · Cited by 24,513
A Unified Approach to Interpreting Model Predictions
Lundberg, S., & Lee, S.-I. (2017) · arXiv · doi:10.48550/arxiv.1705.07874 · Cited by 23,707
"Why Should I Trust You?": Explaining the Predictions of Any Classifier
Ribeiro, M. T., Singh, S., & Guestrin, C. (2016) · Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining · doi:10.1145/2939672.2939778 · Cited by 17,555
Visualizing and understanding convolutional networks
Zeiler, M. D., & Fergus, R. (2014) · Lecture Notes in Computer Science · doi:10.1007/978-3-319-10590-1_53 · Cited by 15,839
A survey on deep learning in medical image analysis
Litjens, G., Kooi, T., Bejnordi, B. E., Setio, A. A. A., Ciompi, F., Ghafoorian, M., van der Laak, J. A. W. M., van Ginneken, B., & Sánchez, C. I. (2017) · Medical Image Analysis · doi:10.1016/j.media.2017.07.005 · Cited by 15,332
Advances in Neural Information Processing Systems 29
Teymur, O., Zygalakis, K., & Calderhead, B. (2016) · Cited by 13,881
Learning Deep Features for Discriminative Localization
Zhou, B., Khosla, A., Lapedriza, A., Oliva, A., & Torralba, A. (2016) · 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2016.319 · Cited by 11,428
Rudin, C. (2019) · Nature Machine Intelligence · doi:10.1038/s42256-019-0048-x · Cited by 10,561
High-performance medicine: the convergence of human and artificial intelligence
Topol, E. J. (2019) · Nature Medicine · doi:10.1038/s41591-018-0300-7 · Cited by 10,450
Barredo Arrieta, A., Díaz-Rodríguez, N., Del Ser, J., Bennetot, A., Tabik, S., Barbado, A., Garcia, S., Gil-Lopez, S., Molina, D., Benjamins, R., et al. (2020) · Information Fusion · doi:10.1016/j.inffus.2019.12.012 · Cited by 10,423
Research works printed but not linked: 2
- Betancur, 2018, Deep learning for prediction of obstructive disease from fast myocardial perfusion SPECT: a multicenter study, JACC: Cardiovasc. Imaging, 11, 1654
- Wierzynski, 2018, The challenges and opportunities of explainable AI, Intel. Com, 12
Cited by
Sorted by citations
Items 301–371 of 371
Chang, Y.-S., Seong, Y., & Yi, S. (2025) · Proceedings of the Human Factors and Ergonomics Society Annual Meeting · doi:10.1177/10711813251369381
Kirubanantham, P., Saranya, A., Revathi, M., Sayeekumar, M., & Anand, M. S. (2025) · Advances in computational intelligence and robotics book series · doi:10.4018/979-8-3373-2787-7.ch015
Unsupervised learning of multi-omics data enables disease risk prediction in the UK BiobankPreprint
Chiara, R., Gräf, J. F., Avelli, M. P., Medina, R. H., Webel, H. E., Ravn, K., & Rasmussen, S. (2025) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.1101/2025.10.02.679853
Kaiser, C. M., Maurits, N. M., & van Vugt, M. K. (2025) · Research Square · doi:10.21203/rs.3.rs-7618908/v1
Enhancing Breast Cancer Detection Through Explainable Federated Learning Models
Pasha, A., Rahman, S. Z., Ahamed, S. S., Kumar, D. N. P., & Manjunatha, R. (2025) · Lecture notes in networks and systems · doi:10.1007/978-3-032-00715-5_28
Haq, H. U., & Khan, M. A. H. (2025) · Clinics and Practice · doi:10.3390/clinpract15110195
The Role of Artificial Intelligence in Smart Healthcare System
Singhal, S., Sharma, A. K., Sharma, A., Chadha, S., & Chauhan, R. (2025) · Studies in computational intelligence · doi:10.1007/978-3-031-99946-8_1
An OSCE-Inspired Framework for Fine-Tuning Multimodal LLMs in Medical Diagnosis
Panagoulias, D. P., Palamidas, A. F., Virvou, M. K., & Tsihrintzis, G. A. (2025) · Procedia Computer Science · doi:10.1016/j.procs.2025.09.225
Mendiratta, S., Mukkelli, V. G., Kayal, E. B., Khanna, P., & Mehndiratta, A. (2025) · Journal of Clinical Monitoring and Computing · doi:10.1007/s10877-025-01387-8
Li, L., Zhou, H., Wei, J., Huang, J., & Lu, D. (2025) · Journal of Biosciences and Medicines · doi:10.4236/jbm.2025.1311034
Matsuda, K., Nakano, J., & Uemura, O. (2025) · Spinal Cord · doi:10.1038/s41393-025-01157-y
Reconstructing Late Quaternary coastal landscapes by a machine-learning framework
Mattei, G., Sorrentino, A., Pappone, G., Ciaramella, A., & Aucelli, P. P. C. (2025) · Scientific Reports · doi:10.1038/s41598-025-31006-0
Dhull, A., Kataria, A., Kalra, M., & Singh, A. (2025) · Springer series in reliability engineering · doi:10.1007/978-3-031-98728-1_15 · Cited by 1
A Hybrid MobileNetV2–Shallow CNN Learning Model for Disease Classification
Kumar, R. H., & Sunitha, G. (2025) · International Conference on Innovative Data Communication Technologies and Application (ICIDCA) · doi:10.1109/icidca66325.2025.11280381
Ethics and Legal Issues in Computational Biomedical Intelligence
Rathod, M. J., Arora, N., Divakaran, S., Agarwal, A., Ballal, S., & Mohapatra, C. K. (2025) · Wiley eBooks · doi:10.1002/9781394386659.ch5
Evaluation of the INCISIVE Services in Cancer Imaging: A Feasibility Study
Zacharias, L., Hesso, I., Kayyali, R., Charalambous, A., Triumbari, E. K. A., Chouvarda, I., Lončar-Turukalo, T., Sykiotis, S., Symvoulidis, C., Gutierrez-Torre, A., et al. (2026) · Seminars in Oncology Nursing · doi:10.1016/j.soncn.2025.152088
Ethical issues with large language models
Tedesco, R. (2026) · Elsevier eBooks · doi:10.1016/b978-0-443-33592-1.00003-6
Retrieving Filter Spectra in CNN for Explainable Sleep Stage Classification
Goerttler, S., Wang, Y., He, F., & Wu, M. (2025) · IEEE Signal Processing in Medicine and Biology Symposium · doi:10.1109/spmb67169.2025.11345374
Kaur, B., & Chen, D. (2025) · International Conference on Cyber-Physical Social Intelligence (ICCSI) · doi:10.1109/cpsi66656.2025.11343977
Explainable AI for Lymphography Prediction: A SHAP-Based Approach
Kumar, A., Si, T., Nandi, U., & Mallik, S. (2025) · International Conference on Internet of Everything, Microwave Engineering, Communication and Networks (IEMECON) · doi:10.1109/iemecon69302.2025.11365698
Unraveling pedestrian fatality patterns: A comparative study with explainable AI
Sulle, M., Mwakalonge, J., Comert, G., Siuhi, S., & Gyimah, N. K. (2026) · Transportation Research Interdisciplinary Perspectives · doi:10.1016/j.trip.2026.101856
Constantine-Macías, A., Tapia-Rosero, A., Añazco, E. V., González, E. V., Miranda, M. J., Yépez-Guerra, L., Peláez, E., & Loayza, F. R. (2026) · Lecture notes in networks and systems · doi:10.1007/978-3-031-98768-7_10
Explainable AI Based Multi-Class Pneumonia Classification and Anomaly Detection from Chest X-ray
Mahfuz, S., Israt, T., Anannya, M., Hosen, M. B., & Mazumder, R. (2025) · IEEE International Conference on Computing, Applications and Systems (COMPAS) · doi:10.1109/compas67506.2025.11381871
Liang, Y., Yin, Y., Zhen, K., Lin, H., Yu, G., & Li, X. (2026) · Neuropsychiatric Disease and Treatment · doi:10.2147/ndt.s582787
Budhbaware, T., Rathored, J., Basheeruddin, M., & Qausain, S. (2025) · DMIHER International Conference on Artificial Intelligence in Healthcare, Education and Industry (IDICAIHEI) · doi:10.1109/idicaihei65991.2025.11378850
Chen, Z., Wang, Y., & Wu, C. Y. (2026) · Seminars in Ultrasound CT and MRI · doi:10.1053/j.sult.2026.03.001
Hu, J., He, K., Luo, J., Xia, R., Zhang, Y., Ma, X., Shao, Z., & Gao, X. (2026) · BMC Medical Imaging · doi:10.1186/s12880-026-02265-8
Yin, J., Zhang, B., & Zeng, M. (2025) · International Conference on Public Health and Data Science (ICPHDS) · doi:10.1145/3789537.3789543
Machine learning studies of drug-induced nephrotoxicity: a scoping review
Ihsan, M., Chang, S.-T., Chan, W.-K., & Chen, H.-Y. (2026) · Therapeutic Advances in Drug Safety · doi:10.1177/20420986261430234
AI Chatbots for Well-Being Adoption Drivers- A Machine Learning Approach
Tavares, J., Yang, Y., & Oliveira, T. (2026) · Procedia Computer Science · doi:10.1016/j.procs.2026.03.112
Asad, K., Farooqui, N. A., & Ahmad, A. (2026) · Advances in computational intelligence and robotics book series · doi:10.4018/979-8-3373-7600-4.ch014
Equity by Design: Integrating Bias Audits and Fairness Mandates Into Healthcare AI Governance
Athukorala, D., Ahmed, K., & Nowrozy, R. (2026) · Lecture notes in computer science · doi:10.1007/978-981-95-6304-3_1
Attai, K. F., Asuquo, D. E., Akputu, O. K., Obot, O. U., Thomas, C., Uzoka, F.-V., Attai, E., Akwaowo, C., & Uzoka, F. (2026) · Information · doi:10.3390/info17050435
Revolutionizing genomic medicine with artificial intelligence and data analytics
Patra, P. P. (2026) · Elsevier eBooks · doi:10.1016/b978-0-443-27638-5.00013-4
Subhashini, K. (2026) · Advances in computational intelligence and robotics book series · doi:10.4018/979-8-3373-7320-1.ch011
Shi, Y., Chun, Y., Zhang, S., Xu, X., Chen, W., Jiao, L., Wang, L., 屠丽萍, & Xu, J. (2026) · Frontiers in Physiology · doi:10.3389/fphys.2026.1835790
Deep Learning-Enabled Medical Image Interpretation for Early and Explainable Disease Identification
Aziz, A. (2026) · International Conference on Smart Electronic Devices and Intelligent Systems (ICSEDIS) · doi:10.1109/icsedis68157.2026.11518048
Bharti, A., Kumar, S., Kumar, N., Islam, M., & Kumar, S. (2026) · Asian Journal of Civil Engineering · doi:10.1007/s42107-026-01696-6
Exploring the Role of Federated Learning and Split Learning in Advancing Healthcare Technology
Mithaguru, & Fernando, A. V. (2026) · Communications in computer and information science · doi:10.1007/978-981-95-7289-2_12 · Cited by 1
Wadhwa, K., Ananya, D., Mittal, H., & Jindal, S. (2026) · AIP conference proceedings · doi:10.1063/5.0329327
Misgar, M. M., & Bhatia, M. P. S. (2026) · Soft Computing · doi:10.1007/s00500-024-09832-7
Interpretable Hybrid Machine Learning for Student Mental Health Screening
Subhash Chandra Mouli, D., Kovuru, D. R., Ganesh, P., Asifa, S., Dharani, K., & Pandurangaraju, B. (2026) · International Conference on Knowledge Engineering and Communication Systems (ICKECS) · doi:10.1109/ickecs70176.2026.11527614
Explainable Kidney Stone Detection from CT Scans: Integrating Grad-CAM with YOLO Model
Safi, F., Kacem, F., Jallouli, W., & Fathallah, H. (2026) · IEEE Middle East and North Africa Communications Conference (MENACOMM) · doi:10.1109/menacomm69507.2026.11532874
Cook, R. M., Dube, A., Asaduzzaman, M., Smith, H., Whitehouse, C. L., Pearce, R., Beales, T., Blackwell, L., Miller, J., Varnals, G., et al. (2026) · International Journal of Nursing Studies · doi:10.1016/j.ijnurstu.2026.105613
Inspection of joint application of smote and auto-encoder for binary classification
Chowdhury, S., Sahoo, A., Panda, B., Masanta, A., Kumar, A. S. P., Nayak, G., & Dutta, P. (2026) · AIP conference proceedings · doi:10.1063/5.0317864
Mahbub, M. J. A., Dipu, S. A., Hasan, R., Islam, M. F. U., Hasan, M. M., Tahsin, A., Alam, M. G. R., & Uddin, M. Z. (2026) · PeerJ Computer Science · doi:10.7717/peerj-cs.3905
Preaching Meaning, Practicing the Motions: A Critical Analysis of XAI Research
van der Linden, I., & Ghebreab, S. (2026) · ACM Conference on Fairness, Accountability, and Transparency (FAccT) · doi:10.1145/3805689.3812356
Patra, S. C., Maheswari, B. U., & Naik, G. R. (2026) · Journal of Computational Design and Engineering · doi:10.1093/jcde/qwag059
Zhang, X., Huang, Y., Jie, W., Zhang, X., Chen, H. Z., Luo, Y., Xie, J., Xu, Z. Y., & Xu, Y. H. (2026) · Journal of Applied Clinical Medical Physics · doi:10.1002/acm2.70667
Enhancing Explainability in AI Models for Blood Glucose Prediction in Diabetes PatientsPreprint
Nguyen, L. N., & Wong, K.-C. (2026) · Research Square · doi:10.21203/rs.3.rs-10117658/v1
Zhang, A., Yu, J., Tao, Q., Zhou, T., Sheng, A., Li, R., Wu, X., & Yu, Z. (2026) · Journal of Medical Internet Research · doi:10.2196/89300
Naser, M. Z. (2026) · Safety Science · doi:10.1016/j.ssci.2026.107367
Venter, M., & Naidoo, K. J. (2026) · Research Square · doi:10.21203/rs.3.rs-10155568/v1
Biosensor-Based Disease Detection System Integrated with Computation Tools
Sharma, S., Verma, S., Saxena, P., Saxena, A., & Kumar, S. (2026) · Springer eBooks · doi:10.1007/978-981-95-9386-6_12
A Novel Optimal Key Generation Technique for AI Driven Health Care Systems
Himanshu, Pandey, B., Bhardwaj, M. K., Singh, A., Sharma, M., & Singh, P. (2026) · Lecture notes in networks and systems · doi:10.1007/978-3-032-28307-8_37
Feature Extraction and Classification of ECG Signals Using Multi-channel 1-D CNN
Barhumi, I., Shakhatreh, W. O. A., Meqbel, O., & Yousif, M. (2026) · Lecture notes in electrical engineering · doi:10.1007/978-3-032-13917-7_4
Seeing How AI Thinks: UMAP-Based Interpretation of Deep Medical Imaging Models
Balaman, C. C. R., Peñanueva, B. J., Lapeceros, L. E. E. B., & Cababat, S. A. (2026) · Communications in computer and information science · doi:10.1007/978-981-92-1546-1_30
The Statistical Assessment of Bayes–“sub”Optimal Binary Machine Learning Classifier Risk
Chan, A., Gashi, I., Gopalakrishnan, S., Pattabiraman, K., & Salako, K. (2026) · Lecture notes in computer science · doi:10.1007/978-3-032-34867-8_19
Li, H., Cai, J., Zhou, X., Zhou, L., Li, Y., & Song, T. (2026) · International Journal of Radiation Oncology*Biology*Physics · doi:10.1016/j.ijrobp.2026.07.060
LIME vs. SHAP: A Quantitative Comparison of Local Explanation Methods for Random Forest Classifiers
Samal, K., Miras, S., & Sultan, K. (2026) · International Conference on Cybersecurity, Digital Forensics, and AI Applications (ICCSDFAI) · doi:10.1109/iccsdfai70505.2026.11648058
Yue, R., Yin, H., Zhang, X., Li, C., Meng, D., Li, J., Zheng, D., Liu, S., & Wang, K. (2026) · Frontiers in Immunology · doi:10.3389/fimmu.2026.1910588
From EEG Source Localization to Brain Network Analysis: A Computational Neuroscience Perspective
Lee, J., & Eom, T. (2026) · Biology · doi:10.3390/biology15181622
A comprehensive review of deep learning methods for weed classification in precision agriculture
Ekene, N. C., Okoye, F. A., Chidi, E. U., & NNENNA, O. M. (2026) · Discover Artificial Intelligence · doi:10.1007/s44163-026-01916-7 · Cited by 1
Vutla, A., & Tiwari, B. K. (2026) · International Conference Computational Vision and Bio Inspired Computing · doi:10.1109/iccvbic71195.2026.11688422
yman, C., Benamri, I., Ferjouchia, H., Ferjouchia, Z., & Rabhi, A. (2026) · Neural Computing and Applications · doi:10.1007/s00521-026-12460-x
ÇAYIRTEPE, Z., & Demir, B. (2026) · Hacettepe sağlık idaresi dergisi · doi:10.61859/hacettepesid.1922869
El-Kholy, A., Abdelrahman, N., & Hassan, K. (2023) · Journal of Health Informatics and Digital Systems · doi:10.68159/p837352065
Mehmood, A., Mehmood, F., & Kim, J. (2026) · Mathematics · doi:10.3390/math14193542
Moldt, J.-A., Wietek, T. F., Nieselt, K., Zabel, S., Claassen, M., Wagner, S., Keim, U., & Herrmann‐Werner, A. (2026) · JMIR Medical Education · doi:10.2196/104151
Boretti, A. A. (2026) · Journal of Evaluation in Clinical Practice · doi:10.1111/jep.70661
Montilla, R. E., Zapata, R. E., & Guerrero, E. J. C. (2025) · SAP Artificial Intelligence in Dentistry · doi:10.62486/aid202524
Metadata
- Authors
- Rasheed, K., Qayyum, A., Ghaly, M., Al-Fuqaha, A., Razi, A., & Qadir, J.
- Affiliations
- Information Technology University (PK)
- Hamad bin Khalifa University (QA)
- Canadian Institute for Advanced Research (CA)
- Qatar University (QA)
- Corresponding author
- Qadir, Junaid
- Venue
- Computers in Biology and Medicine
- Publisher
- Elsevier BV
- Published
- 2022-10
- Volume
- 149
- Pages
- 106043
- ISSN
- 0010-4825 (ISSN-L)
- OpenAlex
- W4294839477
- Cited by
- 372
- References
- 217
- Citations by year
- 2026: 77 · 2025: 134 · 2024: 104 · 2023: 43 · 2022: 11
- FWCI
- 5.10
- Type
- journal article
- Language
- en
- Open access
- hybrid
- doi.org
- Licence
- Locations
- Computers in Biology and Medicine (publishedVersion, cc-by)
- PubMed (publishedVersion)
- Figshare (submittedVersion, cc-by)
- Qatar University QSpace (Qatar University) (submittedVersion)
- Funders
- Qatar National Research Fund
- Qatar National Library
- Qatar Foundation
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Explainable Artificial Intelligence (XAI)
- Health Sciences › Medicine › Health Informatics › Artificial Intelligence in Healthcare and Education
- Physical Sciences › Computer Science › Artificial Intelligence › Machine Learning in Healthcare
- Keywords
- explainable machine learning, model interpretability, health care, trustworthiness, ethical considerations, black-box models, clinical decision support, liability
- MeSH terms
- Machine Learning, Delivery of Health Care, Health Facilities, Humans, Surveys and Questionnaires
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
A language model drew the claims from the paper on 2026-10-09, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).
A second model labelled each claim on 2026-10-09 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).
Cited sources are checked with a language model’s help (gpt-6-sol).
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
