mlsanityData to 14 Oct 2026

Better artificial intelligence does not mean better models of biology

arXiv 2504.16940 ↗First posted 8 Apr 2025q-bio.NC cs.AI cs.CVcited by 10

Not on the map. Why?

Abstract (arXiv)

Deep neural networks (DNNs) once showed increasing alignment with primate perception and neural responses as they improved on vision benchmarks, raising hopes that advances in AI would yield better models of biological vision. However, we show across three benchmarks that this alignment is now plateauing - and in some cases worsening - as DNNs scale to human or superhuman accuracy. This divergence may reflect the adoption of visual strategies that differ from those used by primates. These findings challenge the view that progress in artificial intelligence will naturally translate to neuroscience. We argue that vision science must chart its own course, developing algorithms grounded in biological visual systems rather than optimizing for benchmarks based on internet-scale datasets.

Key facts. Cited by 10 papers in the map (data through 14 Oct 2026). Builds on 69 of its 149 references that are in the map. First posted on arXiv 8 Apr 2025.

Builds on

69 of 149 references.

References · 69

  1. 2017Attention Is All You Needcited by 60,361
  2. 2015Deep Residual Learning for Image Recognitioncited by 52,096
  3. 2021Learning Transferable Visual Models From Natural Language Supervisioncited by 33,744
  4. 2020An Image is Worth 16x16 Words: Transformers for Image Recognition at Scalecited by 30,186
  5. 2014Very Deep Convolutional Networks for Large-Scale Image Recognitioncited by 15,364
  6. 2023DINOv2: Learning Robust Visual Features without Supervisioncited by 7,974
  7. 2020Scaling Laws for Neural Language Modelscited by 6,586
  8. 20233D Gaussian Splatting for Real-Time Radiance Field Renderingcited by 6,226
  9. 2022LAION-5B: An open large-scale dataset for training next generation image-text modelscited by 4,152
  10. 2019Benchmarking Neural Network Robustness to Common Corruptions and Perturbationscited by 2,877
  11. 2013Deep Inside Convolutional Networks: Visualising Image Classification Models and Saliency Mapscited by 2,371
  12. 2019Searching for MobileNetV3cited by 1,994
38 more
  1. 2020Shortcut Learning in Deep Neural Networkscited by 1,969
  2. 2016Identity Mappings in Deep Residual Networkscited by 1,965
  3. 20234D Gaussian Splatting for Real-Time Dynamic Scene Renderingcited by 1,157
  4. 2019Learning Robust Global Representations by Penalizing Local Predictive Powercited by 1,013
  5. 2021Scaling Vision Transformerscited by 999
  6. 2023Scaling Vision Transformers to 22 Billion Parameterscited by 694
  7. 2021ImageNet-21K Pretraining for the Massescited by 613
  8. 2020Scaling Laws for Autoregressive Generative Modelingcited by 522
  9. 2024The Platonic Representation Hypothesiscited by 382
  10. 2023EVA-02: A Visual Representation for Neon Genesiscited by 381
  11. 2018Feature Denoising for Improving Adversarial Robustnesscited by 314
  12. 2023Reasoning or Reciting? Exploring the Capabilities and Limitations of Language Models Through Counterfactual Taskscited by 246
  13. 2018Generalisation in humans and deep neural networkscited by 237
  14. 2021Partial success in closing the gap between human and machine visioncited by 185
  15. 2023Embers of Autoregression: Understanding Large Language Models Through the Problem They are Trained to Solvecited by 178
  16. 2020TResNet: High Performance GPU-Dedicated Architecturecited by 119
  17. 2018Learning long-range spatial dependencies with horizontal gated-recurrent unitscited by 98
  18. 2020Beyond accuracy: quantifying trial-by-trial behaviour of CNNs and humans by measuring error consistencycited by 78
  19. 2022Toward Next-Generation Artificial Intelligence: Catalyzing the NeuroAI Revolutioncited by 78
  20. 2021Look at the Variance! Efficient Black-box Explanations with Sobol-based Sensitivity Analysiscited by 65
  21. 2020Self-supervised learning through the eyes of a childcited by 61
  22. 2022Harmonizing the object recognition strategies of deep neural networks with humanscited by 59
  23. 2019Recurrence is required to capture the representational dynamics of the human visual systemcited by 58
  24. 2016DeepGaze II: Reading fixations from deep features trained on object recognitioncited by 55
  25. 2018Emergence of grid-like representations by training recurrent neural networks to perform spatial localizationcited by 52
  26. 2014Deep Neural Networks Rival the Representation of Primate IT Cortex for Core Visual Object Recognitioncited by 50
  27. 2022The neuroconnectionist research programmecited by 48
  28. 2016Bridging the Gaps Between Residual Learning, Recurrent Neural Networks and Visual Cortexcited by 42
  29. 2023Neither hype nor gloom do DNNs justicecited by 40
  30. 2017Deep Learning Models of the Retinal Response to Natural Scenescited by 36
  31. 2019Disentangling neural mechanisms for perceptual groupingcited by 28
  32. 2021Towards robust vision by multi-task learning on monkey visual cortexcited by 28
  33. 2018Task-Driven Convolutional Recurrent Models of the Visual Systemcited by 27
  34. 2023Unlocking Feature Visualization for Deeper Networks with MAgnitude Constrained Optimizationcited by 27
  35. 2019Deep neuroethology of a virtual rodentcited by 23
  36. 2020Recurrent neural circuits for contour detectioncited by 23
  37. 2023Learning high-level visual representations from a child's perspective without strong inductive biasescited by 22
  38. 2024Dimensions underlying the representational alignment of deep neural networks with humanscited by 22

50 most cited of 69 All 69 (JSON)

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 7; 2026: 3.

10 papers.

  1. 2025To model human linguistic prediction, make LLMs less superhumancited by 11
  2. 2025Deep RL Needs Deep Behavior Analysis: Exploring Implicit Planning by Model-Free Agents in Open-Ended Environmentscited by 4
  3. 2025Vision-Language Models Mistake Head Orientation for Gaze Direction: Nonverbal Conversation Cuescited by 2
  4. 2025From Prediction to Understanding: Will AI Foundation Models Transform Brain Science?cited by 2
  5. 2026Attention Alignment Between Humans and Vision-Language Modelscited by 2
  6. 2025Predicting upcoming visual features during eye movements yields scene representations aligned with human visual cortexcited by 1
  7. 2025Quantifying task-relevant representational similarity using decision variable correlationcited by 0
  8. 2025A Unified Geometric Space for Topological Alignment Between Transformer-Based Models and Human Brain Networkscited by 0
  9. 2026The Umwelt Representation Hypothesis: Rethinking Universalitycited by 0
  10. 2026Cognitive Alignment At No Cost: Inducing Human Attention Biases For Interpretable Vision Transformerscited by 0