The Importance of Understanding Deep Learning
Open accessThe DOI matches its Crossref record; 10 of the 10 references it cites were found in a registry or an index.
Abstract
Some machine learning models, in particular deep neural networks (DNNs), are not very well understood; nevertheless, they are frequently used in science. Does this lack of understanding pose a problem for using DNNs to understand empirical phenomena? Emily Sullivan has recently argued that understanding with DNNs is not limited by our lack of understanding of DNNs themselves. In the present paper, we will argue, contra Sullivan, that our current lack of understanding of DNNs does limit our ability to understand with DNNs. Sullivan’s claim hinges on which notion of understanding is at play. If we employ a weak notion of understanding, then her claim is tenable, but rather weak. If, however, we employ a strong notion of understanding, particularly explanatory understanding, then her claim is not tenable.
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
Lipton, Z. C. (2018) · Queue · doi:10.1145/3236386.3241340 · Cited by 5,427
Causation, Prediction, and Search
Spirtes, P., Glymour, C., & Scheines, R. (2001) · doi:10.7551/mitpress/1754.001.0001 · Cited by 4,521
Understanding deep learning (still) requires rethinking generalization
Zhang, C., Bengio, S., Hardt, M., Recht, B., & Vinyals, O. (2021) · Communications of the ACM · doi:10.1145/3446776 · Cited by 2,400 · cited 2×
“However, the reasons why these DNNs generalize well is only insufficiently understood (Zhang et al., 2017; Shwartz-Ziv & Tishby, 2017; Alain & Bengio, 2016; Zhang et al., 2021; Berner et al., 2021).”
BackgroundIntelligible 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 · cited 1×
“Caruana et al. (2015) provide a notorious example of a model for hospital admission triage that is (at least to a large extent) predictively successful, but makes wrong and harmful predictions for a subset of cases.”
BackgroundSanity Checks for Saliency Maps
Adebayo, J., Gilmer, J., Muelly, M., Goodfellow, I., Hardt, M., & Kim, B. (2018) · arXiv · doi:10.48550/arxiv.1810.03292 · Cited by 1,294
Understanding intermediate layers using linear classifier probes
Alain, G., & Bengio, Y. (2016) · arXiv · doi:10.48550/arxiv.1610.01644 · Cited by 581
Accountability of AI Under the Law: The Role of Explanation
Doshi-Velez, F., Kortz, M., Budish, R., Bavitz, C., Gershman, S. J., O'Brien, D., Scott, K., Shieber, S., Waldo, J., Weinberger, D., et al. (2017) · doi:10.2139/ssrn.3064761 · Cited by 457
Transparency in Algorithmic and Human Decision-Making: Is There a Double Standard?
Zerilli, J., Knott, A., Maclaurin, J., & Gavaghan, C. (2019) · Philosophy & Technology · doi:10.1007/s13347-018-0330-6 · Cited by 420
Solving the Black Box Problem: A Normative Framework for Explainable Artificial Intelligence
Zednik, C. (2021) · Philosophy & Technology · doi:10.1007/s13347-019-00382-7 · Cited by 419
Empiricism without magic: transformational abstraction in deep convolutional neural networks
Buckner, C. (2018) · Synthese · doi:10.1007/s11229-018-01949-1 · Cited by 149 · cited 1×
“For instance, Buckner (2018, 2019) has suggested that we can explain the classifications of some DNNs with so-called “transformational abstraction”.”
Background
Research works printed but not linked: 26
- Baumberger, C., Beisbart, C., & Brun, G. (2017). What is Understanding? An Overview of Recent Debates in Epistemology and Philosophy of Science. In S. G. C. Baumberger and S. Ammon, eds., Explaining Understanding: New Perspectives from Epistemolgy and Philosophy of Science. Routledge, pp. 1–34.
- Berner, J., Grohs, P., Kutyniok, G., & Petersen, P. (2021). Theory of Deep Learning, chap. The Modern Mathematics of Deep Learning: Cambridge University Press.
- Buckner, C. (2019). Deep learning: A philosophical introduction. Philosophy Compass e12625. · DOI · Paper page
- Creel, K. A. (2020). Transparency in complex computational systems. Philos. Sci., 87(4), 568–89. · DOI · Paper page
- de Regt, H. W. (2017). Understanding Scientific Understanding. New York: Oxford university press. · DOI · Paper page
- Fischer, I. (2020). The Conditional Entropy Bottleneck. ArXiv:2002.05379v1. · arXiv
- Gijsbers, V. (2013). Understanding, explanation, and unification. Stud. Hist. Philos. Sci. Part A, 44(3), 516–522. · DOI
- Goodfellow, I., Bengio, Y., & Courville, A. (2016). Deep Learning. MIT press.
- Grüne-Yanoff, T. (2009). The explanatory potential of artificial societies. Synthese, 169, 539–555. · DOI
- Hartmann, S. (1999). Models and Stories in Hadron Physics. In M. S. Morgan and M. Morrison, eds., Models as Mediators. Perspectives on Natural and Social Sciences. Cambridge: Cambridge University Press, pp. 326–346. · doi:10.1017/cbo9780511660108.012
- Hastie, T., Tibshirani, R., & Friedman, J. (2009). The Elements of Statistical Learning. Springer Series in Statistics. Springer, second ed. · DOI · Paper page
- Jebeile, J., Lam, V., & Räz, T. (2021). Understanding climate change with statistical downscaling and machine learning. Synthese, 199, 1877–97. · DOI
- Khalifa, K. (2017). Understanding, Explanation, and Scientific Knowledge. Cambridge: Cambridge University Press. · DOI · Paper page
- Kvanvig, J. L. (2003). The Value of Knowledge and the Pursuit of Understanding. Cambridge university press. · DOI · Paper page
- Nielsen, M. A. (2015). Neural Networks and Deep Learning. Determination Press.
- Pearl, J. (2009). Causality: Models, Reasoning, and Inference. Cambridge: Cambridge university press. · DOI · Paper page
- Räz, T. (2022). Understanding deep learning with statistical relevance. Philos. Sci., 89(1), 20–41. · DOI
- Ribeiro, M. T., Singh, S., & Guestrin, C. (2016). “Why Should I Trust You?” Explaining the Predictions of Any Classifier. ArXiv:1602.04938. · DOI · Paper page
- Rice, C. (2016). Factive scientific understanding without accurate representation. Biol. Philos., 31, 81–102. · DOI
- Rudin, C. (2019). Stop explaining black box machine learning models for high stakes decisions and use interpretable models instead. Nat. Mach. Intell., 1, 206–15. · DOI · Paper page
- Shwartz-Ziv, R., & Tishby, N. (2017). Opening the black box of Deep Neural Networks via Information. ArXiv:1703.00810. · arXiv
- Sullivan, E. (2022). Understanding from machine learning models. Br. J. Philos. Sci., 73(1), 109–133. · DOI · Paper page
- Verreault-Julien, P. (2019). How could models possibly provide how-possibly explanations? Stud. hist. philos. sci. Part A, 73, 22–33. · DOI
- Vidal, R., Bruna, J., Giryes, R., & Soatto, S. (2017). Mathematics of deep learning. ArXiv:1712.04741. · arXiv
- Wilkenfeld, D. A. (2019). Understanding as compression. Philos. Stud., 176, 2807–31. · DOI
- Zhang, C., Bengio, S., Hardt, M., Recht, B., & Vinyals, O. (2017). Understanding Deep Learning Requires Rethinking Generalization. ArXiv:1611.03530. · arXiv
Cited by
Machine learning for the physics of climate
Bracco, A., Brajard, J., Dijkstra, H. A., Hassanzadeh, P., Lessig, C., & Monteleoni, C. (2024) · Nature Reviews Physics · doi:10.1038/s42254-024-00776-3 · Cited by 79
Explainable AI and Causal Understanding: Counterfactual Approaches Considered
Baron, S. (2023) · Minds and Machines · doi:10.1007/s11023-023-09637-x · Cited by 58
Predicting and explaining with machine learning models: Social science as a touchstone
Buchholz, O., & Grote, T. (2023) · Studies in History and Philosophy of Science · doi:10.1016/j.shpsa.2023.10.004 · Cited by 15
Schmidt, E., Putora, P. M., & Fijten, R. (2025) · Philosophy & Technology · doi:10.1007/s13347-024-00834-9 · Cited by 14
Metadata
- Authors
- Räz, T., & Beisbart, C.
- Venue
- Erkenntnis
- Published
- 2022-08-07
- Volume
- 89
- Issue
- 5
- Pages
- 1823-1840
- ISSN
- 0165-0106 (ISSN-L)
- 1572-8420
- OpenAlex
- W4290458740
- Cited by
- 46
- References
- 10
- FWCI
- 40.99
- Type
- journal article
- Language
- en
- Open access
- hybrid
- Licence
- creativecommons.org/licenses/by/4.0 (tdm)
- creativecommons.org/licenses/by/4.0 (vor)
- Funders
- Cogito Foundation
- schweizerischer nationalfonds zur förderung der wissenschaftlichen forschung (197504)
- University of Bern
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
How to cite
Cite CoChat for the checks on this page, and the paper for its own findings.
- This page
- CoChat. (2026). Citation check: The Importance of Understanding Deep Learning. https://app.cochat.ai/papers/10.1007/s10670-022-00605-y
- The paper
- Räz, T., & Beisbart, C. (2024). The Importance of Understanding Deep Learning. Erkenntnis. https://doi.org/10.1007/s10670-022-00605-y
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.