Emergence of brain-like mirror-symmetric viewpoint tuning in convolutional neural networks
- Rockefeller University, United States
- Columbia University, United States
Abstract
Primates can recognize objects despite 3D geometric variations such as in-depth rotations. The computational mechanisms that give rise to such invariances are yet to be fully understood. A curious case of partial invariance occurs in the macaque face-patch AL and in fully connected layers of deep convolutional networks in which neurons respond similarly to mirror-symmetric view (e.g., left and right profiles). Why does this tuning develop? Here, we propose a simple learning-driven explanation for mirror-symmetric viewpoint tuning. We show that mirror-symmetric viewpoint tuning for faces emerges in the fully connected layers of convolutional deep neural networks trained on object recognition tasks, even when the training dataset does not include faces. First, using 3D objects rendered from multiple views as test stimuli, we demonstrate that mirror-symmetric viewpoint tuning in convolutional neural network models is not unique to faces: it emerges for multiple object categories with bilateral symmetry. Second, we show why this invariance emerges in the models. Learning to discriminate among bilaterally symmetric object categories induces reflection-equivariant intermediate representations. AL-like mirror-symmetric tuning is achieved when such equivariant responses are spatially pooled by downstream units with sufficiently large receptive fields. These results explain how mirror-symmetric viewpoint tuning can emerge in neural networks, providing a theory of how they might emerge in the primate brain. Our theory predicts that mirror-symmetric viewpoint tuning can emerge as a consequence of exposure to bilaterally symmetric objects beyond the category of faces, and that it can generalize beyond previously experienced object categories.
Data availability
The stimulus set and the source code required for reproducing our results are available at https://gitfront.io/r/afarzmahdi/p666tmWy7YuY/AL-symmetry-manuscript-codes/.
-
Object categories across viewshttps://github.com/amirfarzmahdi/AL-Symmetry.
Article and author information
Author details
Amirhossein Farzmahdi
Winrich A Freiwald
Nikolaus Kriegeskorte
Funding
National Eye Institute (R01EY021594)
- Winrich A Freiwald
National Eye Institute (R01EY029998)
- Winrich A Freiwald
National Institute of Neurological Disorders and Stroke (RF1NS128897)
- Nikolaus Kriegeskorte
Naval Research Laboratory (N00014-20-1-2292)
- Winrich A Freiwald
Charles H. Revson Foundation
- Tal Golan
The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.
Copyright
© 2024, Farzmahdi et al.
This article is distributed under the terms of the Creative Commons Attribution License permitting unrestricted use and redistribution provided that the original author and source are credited.
Metrics
-
- 799
- views
-
- 136
- downloads
-
- 0
- citations
Views, downloads and citations are aggregated across all versions of this paper published by eLife.
Download links
A two-part list of links to download the article, or parts of the article, in various formats.
Downloads (link to download the article as PDF)
Open citations (links to open the citations from this article in various online reference manager services)
Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)
Emergence of brain-like mirror-symmetric viewpoint tuning in convolutional neural networks
eLife 13:e90256.
https://doi.org/10.7554/eLife.90256
Further reading
-
- Neuroscience
Combining electrophysiological, anatomical and functional brain maps reveals networks of beta neural activity that align with dopamine uptake.
-
- Neuroscience
During rest and sleep, memory traces replay in the brain. The dialogue between brain regions during replay is thought to stabilize labile memory traces for long-term storage. However, because replay is an internally-driven, spontaneous phenomenon, it does not have a ground truth - an external reference that can validate whether a memory has truly been replayed. Instead, replay detection is based on the similarity between the sequential neural activity comprising the replay event and the corresponding template of neural activity generated during active locomotion. If the statistical likelihood of observing such a match by chance is sufficiently low, the candidate replay event is inferred to be replaying that specific memory. However, without the ability to evaluate whether replay detection methods are successfully detecting true events and correctly rejecting non-events, the evaluation and comparison of different replay methods is challenging. To circumvent this problem, we present a new framework for evaluating replay, tested using hippocampal neural recordings from rats exploring two novel linear tracks. Using this two-track paradigm, our framework selects replay events based on their temporal fidelity (sequence-based detection), and evaluates the detection performance using each event's track discriminability, where sequenceless decoding across both tracks is used to quantify whether the track replaying is also the most likely track being reactivated.
