mlsanityData to 7 Oct 2026

The Representational Status of Deep Learning Models

arXiv 2303.12032 ↗First posted 21 Mar 2023cs.AI cs.CY cs.LGcited by 1

Not on the map. Why?

Abstract (arXiv)

This paper aims to clarify the representational status of Deep Learning Models (DLMs). While commonly referred to as 'representations', what this entails is ambiguous due to a conflation of functional and relational conceptions of representation. This paper argues that while DLMs represent their targets in a relational sense, in general, we have no good reason to believe that DLMs encode locally semantically decomposable representations of their targets. That is, the representational capacity these models have is largely global, rather than decomposable into stable, local subrepresentations. This result has immediate implications for explainable AI (XAI) and directs attention toward exploring the global relational nature of deep learning representations and their relationship both to models more generally to understand their potential role in future scientific inquiry.

Key facts. Cited by 1 paper in the map (data through 7 Oct 2026). Builds on 25 of its 105 references that are in the map. First posted on arXiv 21 Mar 2023.

Builds on

25 of 105 references.

References · 25

  1. 2013Intriguing properties of neural networkscited by 5,114
  2. 2013Rich feature hierarchies for accurate object detection and semantic segmentationcited by 2,160
  3. 2020Shortcut Learning in Deep Neural Networkscited by 1,928
  4. 2019Invariant Risk Minimizationcited by 1,832
  5. 2018Sanity Checks for Saliency Mapscited by 1,149
  6. 2019Adversarial Examples Are Not Bugs, They Are Featurescited by 950
  7. 2022In-context Learning and Induction Headscited by 930
  8. 2018Robustness May Be at Odds with Accuracycited by 873
  9. 2014Deep Neural Networks are Easily Fooled: High Confidence Predictions for Unrecognizable Imagescited by 774
  10. 2016The Mythos of Model Interpretabilitycited by 716
  11. 2020Underspecification Presents Challenges for Credibility in Modern Machine Learningcited by 488
  12. 2017Interpretation of Neural Networks is Fragilecited by 443
13 more
  1. 2022Talking About Large Language Modelscited by 174
  2. 2019Towards Explainable Artificial Intelligencecited by 106
  3. 2019Understanding artificial intelligence ethics and safetycited by 68
  4. 2018A Theoretical Explanation for Perplexing Behaviors of Backpropagation-based Visualizationscited by 59
  5. 2020Towards falsifiable interpretability researchcited by 53
  6. 2014Deep Neural Networks Rival the Representation of Primate IT Cortex for Core Visual Object Recognitioncited by 50
  7. 2021Why AI is Harder Than We Thinkcited by 40
  8. 2020A Survey of Deep Learning for Scientific Discoverycited by 36
  9. 2018Humans can decipher adversarial imagescited by 31
  10. 2020Overinterpretation reveals image classification model pathologiescited by 28
  11. 2019Generative models as parsimonious descriptions of sensorimotor loopscited by 11
  12. 2022Deep Learning Opacity in Scientific Discoverycited by 9
  13. 2021A Philosophical Understanding of Representation for Neurosciencecited by 6

Built on by

Cited by year

Citing papers per year: 2025: 1.

1 paper.

  1. 2025DNNs, Dataset Statistics, and Correlation Functionscited by 1