Spring-block theory of feature learning in deep neural networks
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2024
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866909662728683520 |
|---|---|
| author | Shi, Cheng Pan, Liming Dokmanić, Ivan |
| author_facet | Shi, Cheng Pan, Liming Dokmanić, Ivan |
| contents | Feature-learning deep nets progressively collapse data to a regular low-dimensional geometry. How this emerges from the collective action of nonlinearity, noise, learning rate, and other factors, has eluded first-principles theories built from microscopic neuronal dynamics. We exhibit a noise-nonlinearity phase diagram that identifies regimes where shallow or deep layers learn more effectively and propose a macroscopic mechanical theory that reproduces the diagram and links feature learning across layers to generalization. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2407_19353 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Spring-block theory of feature learning in deep neural networks Shi, Cheng Pan, Liming Dokmanić, Ivan Disordered Systems and Neural Networks Statistical Mechanics Machine Learning Feature-learning deep nets progressively collapse data to a regular low-dimensional geometry. How this emerges from the collective action of nonlinearity, noise, learning rate, and other factors, has eluded first-principles theories built from microscopic neuronal dynamics. We exhibit a noise-nonlinearity phase diagram that identifies regimes where shallow or deep layers learn more effectively and propose a macroscopic mechanical theory that reproduces the diagram and links feature learning across layers to generalization. |
| title | Spring-block theory of feature learning in deep neural networks |
| topic | Disordered Systems and Neural Networks Statistical Mechanics Machine Learning |
| url | https://arxiv.org/abs/2407.19353 |