Deep convolutional neural networks are not mechanistic explanations of object recognition
Bibliographic Data
| ID | 10808337 |
|---|---|
| Authors | Bojana Grujičić (0000-0003-2551-5070, Humboldt-Universität zu Berlin, corresponding author) |
| Year | 2024 |
| Volume | 203 |
| Issue | 1 |
| Publication date | 2024-01-12 |
| Peer Reviewed | Yes |
| Open Access | Yes |
| Type | ARTICLE |
| Venue | Synthese (JOURNAL) |
| Journal identifiers | ISSN: 0039-7857 • E-ISSN: 1573-0964 |
| Publisher | Springer Science and Business Media LLC (PUBLISHER) |
| DOI | 10.1007/s11229-023-04461-3 |
| OpenAlex | W4390812791 |
| Language | EN |
| References cited | 83 |
Given the extent of using deep convolutional neural networks to model the mechanism of object recognition, it becomes important to analyse the evidence of their similarity and the explanatory potential of these models. I focus on one frequent method of their comparison—representational similarity analysis, and I argue, first, that it underdetermines these models as how-actually mechanistic explanations. This happens because different similarity measures in this framework pick out different mechanisms across DCNNs and the brain in order to correspond them, and there is no arbitration between them in terms of relevance for object recognition. Second, the reason similarity measures are underdetermining to a large degree stems from the highly idealised nature of these models, which undermines their status as how-possibly mechanistic explanatory models of object recognition as well. Thus, building models with more theoretical consideration and choosing relevant similarity measures may bring us closer to the goal of mechanistic explanation
Cognitive neuroscience of visual object recognition · Cognitive psychology · Cognitive science · Convolutional neural network · Deep learning · Epistemology · Focus (optics) · Image (mathematics) · Machine learning · Mechanism (biology) · Metaphysics · Natural language processing · Object (grammar) · Philosophy of language · Philosophy of science · Relevance (law) · Similarity (geometry) · Adversarial Robustness in Machine Learning · Artificial Intelligence · Cell Image Analysis Techniques · Computer Science · Explainable Artificial Intelligence (XAI · Philosophy · Psychology
Explaining the Brain
Parallel Distributed Processing
ImageNet
Respects for similarity.
ImageNet classification with deep convolutional neural networks
What is a mechanism? Thinking about mechanisms across the sciences
Deep learning
Representational similarity analysis in neuroimaging
Mapping representational mechanisms with deep neural networks
The physics of representation
Empiricism without magic
Minimal models and canonical neural computations
Models and mechanisms in psychological explanation
Mechanisms in psychology
Strategies for Discovering Mechanisms
Mechanisms in Cognitive Psychology
The Explanatory Force of Dynamical and Mathematical Models in Neuroscience
Thinking about Mechanisms
Mechanistic Abstraction
Dynamic models of segregation
Are More Details Better? On the Norms of Completeness for Mechanistic Explanations
Decoding the Brain
What was Hodgkin and Huxley’s Achievement
Deep learning
| Citation velocity | historical |
|---|---|
| Highly cited | No |