Use and usability: concepts of representation in philosophy, neuroscience, cognitive science, and computer science

2026/04/15 by Ben Baker, Richard D. Lange, Andrew Richmond +4 · 1 voice
Computer Science · Neuroscience · Psychology · #Embodied and Extended Cognition #Face Recognition and Perception #Philosophy and Theoretical Science #cs.OH

paper · pdf · doi:10.51628/001c.160037

arxiv published 2026/04/15 · arxiv updated 2026/04/15 · openalex publication_date 2026/04/17 · openalex created_date 2026/04/18 · openalex updated_date 2026/08/01

Abstract

Representations play a central role in the study of both biological and artificial intelligence, as well as philosophy of mind. Across neuroscience, computer science, and philosophy, a recurring theme is that representations not only carry information but should be ``useful'' for or ``usable'' by an agent in some sense. Here, we review how the ``usefulness'' of representations has been conceptualized and how it figures into different conceptions of representation. We identify and explore four aspects of use and usability: representations generally carry information; that information may or may not be useful and it may or may not be encoded in a usable format; and the representations may or may not be used downstream. Building on these four aspects of information and use, we then organize existing perspectives on neural representations into three levels: Representations as Information (Level 1); Representations as Usable (Level 2); and Representations as Used (Level 3). Our account is meant to give readers an appreciation for the diversity of notions of ``neural representation,'' help them navigate the vast and multi-disciplinary literature on the topic, and help them clarify the appropriate notion of representation for their own investigations.

Citations

• What Representational Similarity Measures Imply about Decodable Information
• When does compositional structure yield compositional generalization? A kernel theory
• Soft Matching Distance: A metric on neural representations that captures single-neuron tuning
• Nonlinear Independent Component Analysis for Principled Disentanglement in Unsupervised Deep Learning
• Identifiability Results for Multimodal Contrastive Learning
• Emergent World Representations: Exploring a Sequence Model Trained on a Synthetic Task
• The debate over understanding in AI’s large language models
• The Entangled Brain
• Homomorphism Autoencoder -- Learning Group Structured Representations from Observed Transitions
• Similarity and Matching of Neural Network Representations
• Revisiting Model Stitching to Compare Neural Representations
• Toward Causal Representation Learning
• On the Dangers of Stochastic Parrots
• Usable Information and Evolution of Optimal Representations During Training
• Learning Optimal Representations with the Decodable Information Bottleneck
• A Theory of Usable Information Under Computational Constraints
• When to Trust Your Model: Model-Based Policy Optimization
• Learning Representations by Maximizing Mutual Information Across Views
• Are Disentangled Representations Helpful for Abstract Visual Reasoning?
• Similarity of Neural Network Representations Revisited
• How could models possibly provide how-possibly explanations?
• Interpreting Encoding and Decoding Models
• Interpreting encoding and decoding models
• Life-Long Disentangled Representation Learning with Cross-Domain Latent Homologies
• A Survey on Deep Transfer Learning
• The Lottery Ticket Hypothesis: Finding Sparse, Trainable Neural Networks
• Learning Deep Disentangled Embeddings with the F-Statistic Loss
• Is coding a relevant metaphor for building AI? A commentary on "Is coding a relevant metaphor for the brain?", by Romain Brette
• Emergence of Invariance and Disentanglement in Deep Representations
• Why neurons mix: high dimensionality for higher cognition
• A Probabilistic Theory of Deep Learning
• Deep Learning and the Information Bottleneck Principle
• Understanding image representations by measuring their equivariance and equivalence
• Representation Learning: A Review and New Perspectives
• Representation Learning: A Review and New Perspectives
• Efficient representation as a design principle for neural coding and computation
• The information bottleneck method
• Representation is representation of similarities
• The Extended Mind
• Sparse coding with an overcomplete basis set: A strategy employed by V1?
• Doing without representing?
• Minds, brains, and programs
• Neocognitron: A self-organizing neural network model for a mechanism of pattern recognition unaffected by shift in position
• Is Justified True Belief Knowledge?

Cited by

• What Really Lives in the Swamp? Thought Experiments and the Illustration of Scientific Reasoning

Discussions

• This is finally out! Details below, but basically: we gathered a group of philosophers and cognitive/computer scientists to explain what’s up with neural representation. #philsci #philosophy #philmin [bsky, 4 points, 1 comments]

Related

• What does it mean to represent? Mental representations as falsifiable memory patterns
• The Concept of Representation in ML: Beyond Plato and Aristotle
• Representation Engineering: A Top-Down Approach to AI Transparency
• Operationalising Representation in Natural Language Processing
• A Philosophical Understanding of Representation for Neuroscience