Classification of Explainable Artificial Intelligence Methods through Their Output Formats
Open accessThe DOI matches its Crossref record; 54 of the 54 references it cites were found in a registry or an index.
Abstract
Machine and deep learning have proven their utility to generate data-driven models with high accuracy and precision. However, their non-linear, complex structures are often difficult to interpret. Consequently, many scholars have developed a plethora of methods to explain their functioning and the logic of their inferences. This systematic review aimed to organise these methods into a hierarchical classification system that builds upon and extends existing taxonomies by adding a significant dimension—the output formats. The reviewed scientific papers were retrieved by conducting an initial search on Google Scholar with the keywords “explainable artificial intelligence”; “explainable machine learning”; and “interpretable machine learning”. A subsequent iterative search was carried out by checking the bibliography of these articles. The addition of the dimension of the explanation format makes the proposed classification system a practical tool for scholars, supporting them to select the most suitable type of explanation format for the problem at hand. Given the wide variety of challenges faced by researchers, the existing XAI methods provide several solutions to meet the requirements that differ considerably between the users, problems and application fields of artificial intelligence (AI). The task of identifying the most appropriate explanation can be daunting, thus the need for a classification system that helps with the selection of methods. This work concludes by critically identifying the limitations of the formats of explanations and by providing recommendations and possible future research directions on how to build a more generally applicable XAI method. Future work should be flexible enough to meet the many requirements posed by the widespread use of AI in several fields, and the new regulations.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
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
From local explanations to global understanding with explainable AI for trees
Lundberg, S. M., Erion, G., Chen, H., DeGrave, A., Prutkin, J. M., Nair, B., Katz, R., Himmelfarb, J., Bansal, N., & Lee, S.-I. (2020) · Nature Machine Intelligence · doi:10.1038/s42256-019-0138-9 · Cited by 10,551 · cited 1×
“The authors recently proposed a modified version of SHAP, called TreeExplainer [39], specific for tree-based ML models such as those trained with random forests and gradient boosting.”
MethodExplaining and Harnessing Adversarial Examples
Goodfellow, I. J., Shlens, J., & Szegedy, C. (2014) · arXiv · doi:10.48550/arxiv.1412.6572 · Cited by 10,381
Show, Attend and Tell: Neural Image Caption Generation with Visual Attention
Xu, K., Ba, J., Kiros, R., Cho, K., Courville, A., Salakhutdinov, R., Zemel, R., & Bengio, Y. (2015) · arXiv · doi:10.48550/arxiv.1502.03044 · Cited by 8,248
Peeking Inside the Black-Box: A Survey on Explainable Artificial Intelligence (XAI)
Adadi, A., & Berrada, M. (2018) · IEEE Access · doi:10.1109/access.2018.2870052 · Cited by 6,294
Lipton, Z. C. (2018) · Queue · doi:10.1145/3236386.3241340 · Cited by 5,427
Deep Inside Convolutional Networks: Visualising Image Classification Models and Saliency Maps
Simonyan, K., Vedaldi, A., & Zisserman, A. (2013) · arXiv · doi:10.48550/arxiv.1312.6034 · Cited by 5,370 · cited 2×
“[164] 2014 PH L C P SmoothGrad Smilkov et al.”
BackgroundOn Pixel-Wise Explanations for Non-Linear Classifier Decisions by Layer-Wise Relevance Propagation
Bach, S., Binder, A., Montavon, G., Klauschen, F., Müller, K.-R., & Samek, W. (2015) · PLOS ONE · doi:10.1371/journal.pone.0130140 · Cited by 4,524 · cited 10×
“COLE uses saliency maps methods, such as LRP and DeepLift, to estimate these contributions.”
MethodInfluentialLearning Important Features Through Propagating Activation Differences
Shrikumar, A., Greenside, P., & Kundaje, A. (2017) · arXiv · doi:10.48550/arxiv.1704.02685 · Cited by 2,761
Methods for interpreting and understanding deep neural networks
Montavon, G., Samek, W., & Müller, K.-R. (2018) · Digital Signal Processing · doi:10.1016/j.dsp.2017.10.011 · Cited by 2,665
Explaining prediction models and individual predictions with feature contributions
Štrumbelj, E., & Kononenko, I. (2014) · Knowledge and Information Systems · doi:10.1007/s10115-013-0679-x · Cited by 2,286 · cited 2×
“[55] 2014 G C NC Feature Perturbation Štrumbelj and Kononenko [56] 2014 G C/R NC GSA Cortez and Embrechts [44,45] 2011, 2013 G C/R NC GFA Adler et al.”
BackgroundAnchors: High-Precision Model-Agnostic Explanations
Ribeiro, M. T., Singh, S., & Guestrin, C. (2018) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v32i1.11491 · Cited by 2,162 · cited 4×
“Anchor [28] uses two algorithms to extract IF-THEN rules which highlight the features of an input instance, called “anchors”, which are sufficient for a classifier to make a prediction.”
BackgroundMethodInfluentialGoldstein, A., Kapelner, A., Bleich, J., & Pitkin, E. (2015) · Journal of Computational and Graphical Statistics · doi:10.1080/10618600.2014.907095 · Cited by 1,861
Understanding deep image representations by inverting them
Mahendran, A., & Vedaldi, A. (2015) · 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2015.7299155 · Cited by 1,827
Intelligible Models for HealthCare: Predicting Pneumonia Risk and Hospital 30-day Readmission
Caruana, R., Lou, Y., Gehrke, J., Koch, P., Sturm, M., & Elhadad, N. (2015) · Proceedings of the 21th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining · doi:10.1145/2783258.2788613 · Cited by 1,685
Understanding Black-box Predictions via Influence Functions
Koh, P. W., & Liang, P. (2017) · arXiv · doi:10.48550/arxiv.1703.04730 · Cited by 1,514 · cited 3×
“Four methods, namely Quantitative Input Influence (QII) functions [51]; Gradient Feature Auditing (GFA) [52]; Influence functions [53]; and Monotone Influence Measures [54], utilise influence functions”
BackgroundMethodSmoothGrad: removing noise by adding noise
Smilkov, D., Thorat, N., Kim, B., Viégas, F., & Wattenberg, M. (2017) · arXiv · doi:10.48550/arxiv.1706.03825 · Cited by 1,467 · cited 2×
“[153] 2017 PH L C P SWAF Rajani and Mooney [156] 2017 PH L C P”
BackgroundExplaining nonlinear classification decisions with deep taylor decomposition
Montavon, G., Lapuschkin, S., Binder, A., Samek, W., & Müller, K.-R. (2017) · Pattern Recognition · doi:10.1016/j.patcog.2016.11.008 · Cited by 1,357
Kim, B., Wattenberg, M., Gilmer, J., Cai, C., Wexler, J., Viegas, F., & Sayres, R. (2017) · arXiv · doi:10.48550/arxiv.1711.11279 · Cited by 1,240
Explaining machine learning classifiers through diverse counterfactual explanations
Mothilal, R. K., Sharma, A., & Tan, C. (2020) · Proceedings of the 2020 Conference on Fairness, Accountability, and Transparency · doi:10.1145/3351095.3372850 · Cited by 1,233 · cited 2×
“Adversarial examples are instead used by C-CHVAE [220], Diverse Counterfactual Explanations (DiCE) [221], Evasion-Prone Samples Selection [222], Maximum Mean Discrepancy (MMD)-critic [223] and Pertinent Negatives [224].”
BackgroundMethodExplainable artificial intelligence: A survey
Dosilovic, F. K., Brcic, M., & Hlupic, N. (2018) · 2018 41st International Convention on Information and Communication Technology, Electronics and Microelectronics (MIPRO) · doi:10.23919/mipro.2018.8400040 · Cited by 1,147
Interpretable Explanations of Black Boxes by Meaningful Perturbation
Fong, R. C., & Vedaldi, A. (2017) · 2017 IEEE International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2017.371 · Cited by 1,097
Unmasking Clever Hans predictors and assessing what machines really learn
Lapuschkin, S., Wäldchen, S., Binder, A., Montavon, G., Samek, W., & Müller, K.-R. (2019) · Nature Communications · doi:10.1038/s41467-019-08987-4 · Cited by 1,047
Visualizing and Understanding Recurrent Networks
Karpathy, A., Johnson, J., & Fei-Fei, L. (2015) · arXiv · doi:10.48550/arxiv.1506.02078 · Cited by 845 · cited 3×
“Recurrent neural networks: Shifting from pictorial to textual inputs, Cell Activation Values [202] is a method of explainability for LSTMs.”
BackgroundMethodInfluentialInterpretable Convolutional Neural Networks
Zhang, Q., Wu, Y. N., & Zhu, S.-C. (2018) · 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2018.00920 · Cited by 833
Rationalizing Neural Predictions
Lei, T., Barzilay, R., & Jaakkola, T. (2016) · Proceedings of the 2016 Conference on Empirical Methods in Natural Language Processing · doi:10.18653/v1/d16-1011 · Cited by 777 · cited 2×
“[128] 2016 L C T Relevance and Discriminative Loss Hendricks et al.”
BackgroundLetham, B., Rudin, C., McCormick, T. H., & Madigan, D. (2015) · The Annals of Applied Statistics · doi:10.1214/15-aoas848 · Cited by 766 · cited 1×
“systems transparent by construction: (a) Bayesian rule lists [102]; and (b) fuzzy inference systems [117].”
MethodDatta, A., Sen, S., & Zick, Y. (2016) · 2016 IEEE Symposium on Security and Privacy (SP) · doi:10.1109/sp.2016.42 · Cited by 728 · cited 2×
“Four methods, namely Quantitative Input Influence (QII) functions [51]; Gradient Feature Auditing (GFA) [52]; Influence functions [53]; and Monotone Influence Measures [54], utilise influence functions”
BackgroundMethodNotions of explainability and evaluation approaches for explainable artificial intelligence
Vilone, G., & Longo, L. (2021) · Information Fusion · doi:10.1016/j.inffus.2021.05.009 · Cited by 598 · cited 1×
“There are several other factors relative to the quality and effectiveness of an explanation itself that cannot be ignored [8].”
BackgroundSelf-citationThe Building Blocks of Interpretability
Olah, C., Satyanarayan, A., Johnson, I., Carter, S., Schubert, L., Ye, K., & Mordvintsev, A. (2018) · Distill · doi:10.23915/distill.00010 · Cited by 591 · cited 2×
“For instance, [196] used saliency maps as the building blocks of such interfaces to explain the inferential logic of CNNs.”
BackgroundMethodUnderstanding intermediate layers using linear classifier probes
Alain, G., & Bengio, Y. (2016) · arXiv · doi:10.48550/arxiv.1610.01644 · Cited by 581
Accurate intelligible models with pairwise interactions
Lou, Y., Caruana, R., Gehrke, J., & Hooker, G. (2013) · Proceedings of the 19th ACM SIGKDD international conference on Knowledge discovery and data mining · doi:10.1145/2487575.2487579 · Cited by 564 · cited 1×
“Generalised Additive Models [17] and their extension with pairwise interactions (GA2Ms) [18,236] are linear combinations of simple models, called “shape functions”, trained on a single feature (GAMs) or up to two features (GA2Ms).”
MethodIntelligible models for classification and regression
Lou, Y., Caruana, R., & Gehrke, J. (2012) · Proceedings of the 18th ACM SIGKDD international conference on Knowledge discovery and data mining · doi:10.1145/2339530.2339556 · Cited by 547 · cited 2×
“Oblique Treed Sparse Additive Models (OT-SpAMs) [66] divide the input feature spaces into regions with sparse oblique tree splitting and assign local sparse additive predictive models to individual regions.”
MethodGenerating Visual Explanations
Hendricks, L. A., Akata, Z., Rohrbach, M., Donahue, J., Schiele, B., & Darrell, T. (2016) · Lecture Notes in Computer Science · doi:10.1007/978-3-319-46493-0_1 · Cited by 534
Shortliffe, E. H., Davis, R., Axline, S. G., Buchanan, B. G., Green, C. C., & Cohen, S. N. (1975) · Computers and Biomedical Research · doi:10.1016/0010-4809(75)90009-9 · Cited by 493 · cited 2×
“[130] 1975 Rule-based system L C NC Scenarios Vlek et al.”
MethodVisualizing Deep Neural Network Decisions: Prediction Difference Analysis
Zintgraf, L. M., Cohen, T. S., Adel, T., & Welling, M. (2017) · arXiv · doi:10.48550/arxiv.1702.04595 · Cited by 487
Distilling a neural network into a soft decision tree
Frosst, N., & Hinton, G. (2017) · arXiv · doi:10.48550/arxiv.1711.09784 · Cited by 407 · cited 3×
“[86,87] proposed two algorithms that extract DTs from the weights of a DNN.”
BackgroundMethodTowards Automatic Concept-based Explanations
Ghorbani, A., Wexler, J., Zou, J., & Kim, B. (2019) · arXiv · doi:10.48550/arxiv.1902.03129 · Cited by 378
Explaining Recurrent Neural Network Predictions in Sentiment Analysis
Arras, L., Montavon, G., Müller, K.-R., & Samek, W. (2017) · Proceedings of the 8th Workshop on Computational Approaches to Subjectivity, Sentiment and Social Media Analysis · doi:10.18653/v1/w17-5221 · Cited by 361
Factual and Counterfactual Explanations for Black Box Decision Making
Guidotti, R., Monreale, A., Giannotti, F., Pedreschi, D., Ruggieri, S., & Turini, F. (2019) · IEEE Intelligent Systems · doi:10.1109/mis.2019.2957223 · Cited by 349 · cited 2×
“A genetic algorithm is also employed in GLocalX [71,72] to generate local rules that explain the prediction made by a classifier on a specific instance.”
BackgroundMethodRaghu, M., Gilmer, J., Yosinski, J., & Sohl-Dickstein, J. (2017) · arXiv · doi:10.48550/arxiv.1706.05806 · Cited by 335
Auditing black-box models for indirect influence
Adler, P., Falk, C., Friedler, S. A., Nix, T., Rybeck, G., Scheidegger, C., Smith, B., & Venkatasubramanian, S. (2018) · Knowledge and Information Systems · doi:10.1007/s10115-017-1116-3 · Cited by 280 · cited 4×
“Four methods, namely Quantitative Input Influence (QII) functions [51]; Gradient Feature Auditing (GFA) [52]; Influence functions [53]; and Monotone Influence Measures [54], utilise influence functions”
BackgroundMethodInfluentialBeyond Sparsity: Tree Regularization of Deep Models for Interpretability
Wu, M., Hughes, M., Parbhoo, S., Zazzi, M., Roth, V., & Doshi-Velez, F. (2018) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v32i1.11501 · Cited by 213
Explainable decision forest: Transforming a decision forest into an interpretable tree
Sagi, O., & Rokach, L. (2020) · Information Fusion · doi:10.1016/j.inffus.2020.03.013 · Cited by 189
Hendricks, L. A., Hu, R., Darrell, T., & Akata, Z. (2018) · Lecture notes in computer science · doi:10.1007/978-3-030-01216-8_17 · Cited by 171
Learning Model-Agnostic Counterfactual Explanations for Tabular Data
Pawelczyk, M., Broelemann, K., & Kasneci, G. (2020) · Proceedings of The Web Conference 2020 · doi:10.1145/3366423.3380087 · Cited by 167 · cited 3×
“C-CHVAE and DiCE perturb the input features to identify faithful counterfactuals, meaning that they do not represent local outliers and are “connected to regions with substantial data density”, and are diverse enough to approximate local decision boundaries.”
BackgroundMethodGLocalX—From Local to Global Explanations of Black Box AI Models
Setzu, M., Guidotti, R., Monreale, A., Turini, F., Pedreschi, D., & Giannotti, F. (2021) · Artificial Intelligence · doi:10.1016/j.artint.2021.103457 · Cited by 140
A peek into the black box: exploring classifiers by randomization
Henelius, A., Puolamäki, K., Boström, H., Asker, L., & Papapetrou, P. (2014) · Data Mining and Knowledge Discovery · doi:10.1007/s10618-014-0368-8 · Cited by 98
Exact and Consistent Interpretation for Piecewise Linear Neural Networks: A Closed Form Solution
Chu, L., Hu, X., Hu, J., Wang, L., & Pei, J. (2018) · Proceedings of the 24th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining · doi:10.1145/3219819.3220063 · Cited by 87
PALM: Machine Learning Explanations For Iterative Debugging
Krishnan, S., & Wu, E. (2017) · Proceedings of the 2nd Workshop on Human-In-the-Loop Data Analytics · doi:10.1145/3077257.3077271 · Cited by 81
Generating Visual Explanations
Hendricks, L. A., Akata, Z., Rohrbach, M., Donahue, J., Schiele, B., & Darrell, T. (2016) · arXiv · doi:10.48550/arxiv.1603.08507 · Cited by 64 · cited 1×
“Relevance and discriminative loss [25,129] generates textual explanations for an image classifier such as “The bird in the photo is a White Pelican because.”
BackgroundHow to Explain Individual Classification Decisions
(2010) · Journal of Machine Learning Research · doi:10.5555/1756006.1859912 · Cited by 54 · cited 9×
“Sensitivity Analysis [140] generates explanations that correspond to local gradients indicating which features of a sample must be modified to change its predicted label.”
BackgroundInfluentialSupersparse Linear Integer Models for Interpretable Classification
Ustun, B., Tracà, S., & Rudin, C. (2013) · arXiv · doi:10.48550/arxiv.1306.6677 · Cited by 26
Cited by
Longo, L., Brcic, M., Cabitza, F., Choi, J., Confalonieri, R., Ser, J. D., Guidotti, R., Hayashi, Y., Herrera, F., Holzinger, A., et al. (2024) · Information Fusion · doi:10.1016/j.inffus.2024.102301 · Cited by 626
Self-citationA., S., & R., S. (2023) · Decision Analytics Journal · doi:10.1016/j.dajour.2023.100230 · Cited by 432
A Review of Taxonomies of Explainable Artificial Intelligence (XAI) Methods
Speith, T. (2022) · 2022 ACM Conference on Fairness Accountability and Transparency · doi:10.1145/3531146.3534639 · Cited by 337
Explainable Artificial Intelligence in CyberSecurity: A Survey
Capuano, N., Fenza, G., Loia, V., & Stanzione, C. (2022) · IEEE Access · doi:10.1109/access.2022.3204171 · Cited by 330
“389 The stages are ante-hoc and post-hoc, according to Vilone 390 and Longo [31], [32].”
MethodIslam, M. R., Ahmed, M. U., Barua, S., & Begum, S. (2022) · Applied Sciences · doi:10.3390/app12031353 · Cited by 311
Metadata
- Authors
- Vilone, G., & Longo, L.
- Venue
- Machine Learning and Knowledge Extraction
- Publisher
- MDPI AG
- Published
- 2021-08-04
- Volume
- 3
- Issue
- 3
- Pages
- 615-661
- ISSN
- 2504-4990 (ISSN-L)
- OpenAlex
- W3187293469
- Cited by
- 152
- References
- 54
- FWCI
- 8.12
- Type
- journal article
- Language
- en
- Open access
- gold
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.