Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2507.15798 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866912494999568384 |
|---|---|
| author | Hollard, Lilian Mohimont, Lucas Gaveau, Nathalie Steffenel, Luiz-Angelo |
| author_facet | Hollard, Lilian Mohimont, Lucas Gaveau, Nathalie Steffenel, Luiz-Angelo |
| contents | The paper investigates the performance of state-of-the-art low-parameter deep neural networks for computer vision, focusing on bottleneck architectures and their behavior using superlinear activation functions. We address interference in feature maps, a phenomenon associated with superposition, where neurons simultaneously encode multiple characteristics. Our research suggests that limiting interference can enhance scaling and accuracy in very low-scaled networks (under 1.5M parameters). We identify key design elements that reduce interference by examining various bottleneck architectures, leading to a more efficient neural network. Consequently, we propose a proof-of-concept architecture named NoDepth Bottleneck built on mechanistic insights from our experiments, demonstrating robust scaling accuracy on the ImageNet dataset. These findings contribute to more efficient and scalable neural networks for the low-parameter range and advance the understanding of bottlenecks in computer vision. https://caiac.pubpub.org/pub/3dh6rsel |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2507_15798 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Exploring Superposition and Interference in State-of-the-Art Low-Parameter Vision Models Hollard, Lilian Mohimont, Lucas Gaveau, Nathalie Steffenel, Luiz-Angelo Computer Vision and Pattern Recognition The paper investigates the performance of state-of-the-art low-parameter deep neural networks for computer vision, focusing on bottleneck architectures and their behavior using superlinear activation functions. We address interference in feature maps, a phenomenon associated with superposition, where neurons simultaneously encode multiple characteristics. Our research suggests that limiting interference can enhance scaling and accuracy in very low-scaled networks (under 1.5M parameters). We identify key design elements that reduce interference by examining various bottleneck architectures, leading to a more efficient neural network. Consequently, we propose a proof-of-concept architecture named NoDepth Bottleneck built on mechanistic insights from our experiments, demonstrating robust scaling accuracy on the ImageNet dataset. These findings contribute to more efficient and scalable neural networks for the low-parameter range and advance the understanding of bottlenecks in computer vision. https://caiac.pubpub.org/pub/3dh6rsel |
| title | Exploring Superposition and Interference in State-of-the-Art Low-Parameter Vision Models |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2507.15798 |