Machine Learning Interpretability: A Survey on Methods and Metrics
Open accessThe DOI matches its Crossref record; 105 of the 106 references it cites were found in a registry or an index.
Abstract
Machine learning systems are becoming increasingly ubiquitous. These systems’s adoption has been expanding, accelerating the shift towards a more algorithmic society, meaning that algorithmically informed decisions have greater potential for significant social impact. However, most of these accurate decision support systems remain complex black boxes, meaning their internal logic and inner workings are hidden to the user and even experts cannot fully understand the rationale behind their predictions. Moreover, new regulations and highly regulated domains have made the audit and verifiability of decisions mandatory, increasing the demand for the ability to question, understand, and trust machine learning systems, for which interpretability is indispensable. The research community has recognized this interpretability problem and focused on developing both interpretable models and explanation methods over the past few years. However, the emergence of these methods shows there is no consensus on how to assess the explanation quality. Which are the most suitable metrics to assess the quality of an explanation? The aim of this article is to provide a review of the current state of the research field on machine learning interpretability while focusing on the societal impact and on the developed methods and metrics. Furthermore, a complete literature review is presented in order to identify future directions of work on this field.
Overview
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
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
“Transparency enables AI systems to be tested, audited, and debugged, which is a path toward increasing their safety (Carvalho et al., 2019)15.”
From Transparency of AI in Healthcare as a Multilayered System of Accountabilities: Between Legal Requirements and Technical Limitations (2022) · cites this paper as Carvalho et al., 2019
In the citing paper
“In this paper, we follow scholars who view “interpretability” as the broader concept that, inter alia, includes “explainability.” Specifically, we agree with the statement of M. Honneger that “interpretability is the end-goal we want to achieve, and explanations are the tools to reach interpretability” (Honegger, 2018, citation in Carvalho et al., 2019).”
From Transparency of AI in Healthcare as a Multilayered System of Accountabilities: Between Legal Requirements and Technical Limitations (2022) · cites this paper as Carvalho et al., 2019
In the citing paper
“These methods can capture complex, non-linear interactions among multiple factors that traditional models may miss [29,30].”
From Machine learning-driven risk prediction for post-hospitalization diabetes case management: Integrating clinical and social determinants of health (2026) · cites this paper as [29]
In the citing paper
“The required auditability property of Trustworthy AI systems demands the development of practical tools [109] that are capable of verifying desirable properties of neural networks such as stability, sensitivity, relevance or reachability [110], as well as metrics beyond explainability [111-115], such as on traceability, data quality and integrity.”
From Connecting the dots in trustworthy Artificial Intelligence: From AI principles, ethics, and key requirements to responsible AI systems and regulation (2023) · cites this paper as [111]
In the citing paper
“Model performance metrics (e.g., model accuracy) do not always exhibit true prediction decisions [9].”
From Explainable AI (XAI): Core Ideas, Techniques, and Solutions (2023) · cites this paper as [9]
3 more claims about this paper
Create a free account to see every claim other papers make about this one.
Citations
Cites
Deep Residual Learning for Image Recognition
He, K., Zhang, X., Ren, S., & Sun, J. (2016) · 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2016.90 · Cited by 235,474 · cited 1×
“For example, introduced in 2016, deep residual networks (ResNets) [1] are over 200-layers deep and have been shown to beat human-level performance on object recognition tasks.”
Background(2008) · Journal of Machine Learning Research · Cited by 33,051
Greedy function approximation: A gradient boosting machine
Friedman, J. H. (2001) · The Annals of Statistics · doi:10.1214/aos/1013203451 · Cited by 30,731 · cited 1×
“Partial Dependence Plot [142] Global Feature summary Individual Condition Expectation [143] Global/Local Feature summary Accumulated Local Effects Plot [144] Global Feature summary Feature Interaction [145] Global Feature summary Feature Importance [146] Global/Local Feature summary Local Surrogate Model [98] Local Su…”
BackgroundGrad-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 · cited 3×
“Shapley Values is sometimes called by SHAP.”
Background"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
Algorithm AS 136: A K-Means Clustering Algorithm
Hartigan, J. A., & Wong, M. A. (1979) · Applied Statistics · doi:10.2307/2346830 · Cited by 14,523 · cited 1×
“These techniques range from classic descriptive statistics to data visualization methods, including Principal Component Analysis (PCA) [101] and t-SNE [102] (t-Distributed Stochastic Neighbor Embedding), and clustering methods, such as k-means [103] and MMD-critic [96] (Maximum Mean Discrepancy).”
MethodDistilling the Knowledge in a Neural Network
Hinton, G., Vinyals, O., & Dean, J. (2015) · arXiv · doi:10.48550/arxiv.1503.02531 · Cited by 14,204 · cited 1×
“Research in this type of methods exists for some years [134] but recently increased along with the ML interpretability field [135–137].”
MethodDavid, F. N., & Tukey, J. W. (1977) · Biometrics · doi:10.2307/2529486 · Cited by 12,708
Rudin, C. (2019) · Nature Machine Intelligence · doi:10.1038/s42256-019-0048-x · Cited by 10,561
Research works printed but not linked: 24
- Tractica (2019, January 22). Artificial Intelligence Software Market to Reach $105.8 Billion in Annual Worldwide Revenue by 2025. Available online: https://www.tractica.com/newsroom/press-releases/artificial-intelligence-software-market-to-reach-105-8-billion-in-annual-worldwide-revenue-by-2025/.
- Du, M., Liu, N., and Hu, X. (2018). Techniques for Interpretable Machine Learning. arXiv.
- Committee on Technology National Science and Technology Council and Penny Hill Press (2016). Preparing for the Future of Artificial Intelligence, CreateSpace Independent Publishing Platform.
- ACM US Public Council (2019, January 22). Statement on Algorithmic Transparency and Accountability. Available online: https://www.acm.org/binaries/content/assets/public-policy/2017_usacm_statement_algorithms.pdf.
- Kim, B., Malioutov, D.M., and Varshney, K.R. (2016). Proceedings of the 2016 ICML Workshop on Human Interpretability in Machine Learning (WHI 2016). arXiv.
- Kim, B., Malioutov, D.M., Varshney, K.R., and Weller, A. (2017). Proceedings of the 2017 ICML Workshop on Human Interpretability in Machine Learning (WHI 2017). arXiv.
- Kim, B., Varshney, K.R., and Weller, A. (2018). Proceedings of the 2018 ICML Workshop on Human Interpretability in Machine Learning (WHI 2018). arXiv.
- Wilson, A.G., Kim, B., and Herlands, W. (2016). Proceedings of NIPS 2016 Workshop on Interpretable Machine Learning for Complex Systems. arXiv.
- Caruana, R., Herlands, W., Simard, P., Wilson, A.G., and Yosinski, J. (2017). Proceedings of NIPS 2017 Symposium on Interpretable Machine Learning. arXiv.
- Pereira-Fariña, M., and Reed, C. (2017). Proceedings of the 1st Workshop on Explainable Computational Intelligence (XCI 2017), Association for Computational Linguistics (ACL).
- IJCAI (2019, July 12). IJCAI 2018—Workshop on Explainable Artificial Intelligence (XAI). Available online: http://home.earthlink.net/~dwaha/research/meetings/faim18-xai/.
- CD-MAKE (2019, July 12). CD-MAKE 2019 Workshop on explainable Artificial Intelligence. Available online: https://cd-make.net/special-sessions/make-explainable-ai/.
- Lim, B., Smith, A., and Stumpf, S. (2018). ExSS 2018: Workshop on Explainable Smart Systems. CEUR Workshop Proceedings, City, University of London Institutional Repository. Available online: http://openaccess.city.ac.uk/20037/.
- ICAPS (2019, July 12). ICAPS 2018—Workshop on Explainable AI Planning (XAIP). Available online: http://icaps18.icaps-conference.org/xaip/.
- ICAPS (2019, July 12). ICAPS 2019—Workshop on Explainable AI Planning (XAIP). Available online: https://kcl-planning.github.io/XAIP-Workshops/ICAPS_2019.
- CVPR (2019, July 12). CVPR 19—Workshop on Explainable AI. Available online: https://explainai.net/.
- Lipton, Z.C. (2016). The mythos of model interpretability. arXiv.
- Holzinger, A., Biemann, C., Pattichis, C.S., and Kell, D.B. (2017). What do we need to build explainable AI systems for the medical domain?. arXiv.
- Kim, T.W. (2018). Explainable Artificial Intelligence (XAI), the goodness criteria and the grasp-ability test. arXiv.
- Ross, A., Lage, I., and Doshi-Velez, F. (2017, January 4–9). The neural lasso: Local linear sparsity for interpretable explanations. Proceedings of the Workshop on Transparent and Interpretable Machine Learning in Safety Critical Environments, 31st Conference on Neural Information Processing Systems, Long Beach, CA, USA.
- Lee, M., He, X., Yih, W.t., Gao, J., Deng, L., and Smolensky, P. (2015). Reasoning in vector space: An exploratory study of question answering. arXiv.
- Xu, K., Park, D.H., Yi, C., and Sutton, C. (2018). Interpreting Deep Classifier by Visual Distillation of Dark Knowledge. arXiv.
- Koh, P.W., and Liang, P. (2017). Understanding black-box predictions via influence functions. arXiv.
- Bonnans, J.F., and Shapiro, A. (2013). Perturbation Analysis of Optimization Problems, Springer Science & Business Media.
Cited by
Ali, S., Abuhmed, T., El-Sappagh, S., Muhammad, K., Alonso-Moral, J. M., Confalonieri, R., Guidotti, R., Del Ser, J., Díaz-Rodríguez, N., & Herrera, F. (2023) · Information Fusion · doi:10.1016/j.inffus.2023.101805 · Cited by 1,762
Zhai, C., Wibowo, S., & Li, L. D. (2024) · Smart Learning Environments · doi:10.1186/s40561-024-00316-7 · Cited by 1,567
“According to Carvalho et al. (2019), such transparency is pivotal for enhancing understanding of AI dialogue systems, laying a solid foundation for the challenges and advantages of algorithmic decisions, and ensuring that decision-making processes are both accountable and fair.”
BackgroundInfluentialExplainable AI (XAI): Core Ideas, Techniques, and Solutions
Dwivedi, R., Dave, D., Naik, H., Singhal, S., Omer, R., Patel, P., Qian, B., Wen, Z., Shah, T., Morgan, G., et al. (2023) · ACM Computing Surveys · doi:10.1145/3561048 · Cited by 1,386
“, model accuracy) do not always exhibit true prediction decisions [9].”
BackgroundA comprehensive review on ensemble deep learning: Opportunities and challenges
Mohammed, A., & Kora, R. (2023) · Journal of King Saud University - Computer and Information Sciences · doi:10.1016/j.jksuci.2023.01.014 · Cited by 1,127
“Interpretabil-ity (Carvalho et al., 2019) refers to the ability of a user to understand the ensemble outcomes.”
Explainable artificial intelligence: a comprehensive review
Minh, D., Wang, H. X., Li, Y. F., & Nguyen, T. N. (2022) · Artificial Intelligence Review · doi:10.1007/s10462-021-10088-y · Cited by 983
Metadata
- Authors
- Carvalho, D. V., Pereira, E. M., & Cardoso, J. S.
- Affiliations
- Deloitte Portugal, Manuel Bandeira Street, 43, 4150-479 Porto, Portugal
- Deloitte Portugal (Portugal) (PT)
- Universidade do Porto (PT)
- Venue
- Electronics
- Publisher
- MDPI AG
- Published
- 2019-07-26
- Volume
- 8
- Issue
- 8
- Pages
- 832
- ISSN
- 2079-9292 (ISSN-L)
- OpenAlex
- W2964303497
- Cited by
- 1,842
- References
- 106
- Citations by year
- 2026: 224 · 2025: 354 · 2024: 353 · 2023: 313 · 2022: 285 · 2021: 218 · 2020: 91 · 2019: 3
- FWCI
- 68.32
- Type
- journal article
- Language
- en
- Open access
- gold
- dx.doi.org
- Licence
- Locations
- Electronics (publishedVersion, cc-by)
- DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Explainable Artificial Intelligence (XAI)
- Physical Sciences › Computer Science › Software › Software Testing and Debugging Techniques
- Physical Sciences › Computer Science › Artificial Intelligence › Ethics and Social Impacts of AI
- Keywords
- model interpretability, explainable AI, explanation quality metrics, interpretable models, explanation methods, black-box models
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.
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
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.