Saved in:
| Main Authors: | , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2023
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2302.05440 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866917619474366464 |
|---|---|
| author | Srinivasan, Ravi Mignacco, Francesca Sorbaro, Martino Refinetti, Maria Cooper, Avi Kreiman, Gabriel Dellaferrera, Giorgia |
| author_facet | Srinivasan, Ravi Mignacco, Francesca Sorbaro, Martino Refinetti, Maria Cooper, Avi Kreiman, Gabriel Dellaferrera, Giorgia |
| contents | "Forward-only" algorithms, which train neural networks while avoiding a backward pass, have recently gained attention as a way of solving the biologically unrealistic aspects of backpropagation. Here, we first address compelling challenges related to the "forward-only" rules, which include reducing the performance gap with backpropagation and providing an analytical understanding of their dynamics. To this end, we show that the forward-only algorithm with top-down feedback is well-approximated by an "adaptive-feedback-alignment" algorithm, and we analytically track its performance during learning in a prototype high-dimensional setting. Then, we compare different versions of forward-only algorithms, focusing on the Forward-Forward and PEPITA frameworks, and we show that they share the same learning principles. Overall, our work unveils the connections between three key neuro-inspired learning rules, providing a link between "forward-only" algorithms, i.e., Forward-Forward and PEPITA, and an approximation of backpropagation, i.e., Feedback Alignment. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2302_05440 |
| institution | arXiv |
| publishDate | 2023 |
| record_format | arxiv |
| spellingShingle | Forward Learning with Top-Down Feedback: Empirical and Analytical Characterization Srinivasan, Ravi Mignacco, Francesca Sorbaro, Martino Refinetti, Maria Cooper, Avi Kreiman, Gabriel Dellaferrera, Giorgia Machine Learning "Forward-only" algorithms, which train neural networks while avoiding a backward pass, have recently gained attention as a way of solving the biologically unrealistic aspects of backpropagation. Here, we first address compelling challenges related to the "forward-only" rules, which include reducing the performance gap with backpropagation and providing an analytical understanding of their dynamics. To this end, we show that the forward-only algorithm with top-down feedback is well-approximated by an "adaptive-feedback-alignment" algorithm, and we analytically track its performance during learning in a prototype high-dimensional setting. Then, we compare different versions of forward-only algorithms, focusing on the Forward-Forward and PEPITA frameworks, and we show that they share the same learning principles. Overall, our work unveils the connections between three key neuro-inspired learning rules, providing a link between "forward-only" algorithms, i.e., Forward-Forward and PEPITA, and an approximation of backpropagation, i.e., Feedback Alignment. |
| title | Forward Learning with Top-Down Feedback: Empirical and Analytical Characterization |
| topic | Machine Learning |
| url | https://arxiv.org/abs/2302.05440 |