JPC: Flexible Inference for Predictive Coding Networks in JAX
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2024
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866929616726261760 |
|---|---|
| author | Innocenti, Francesco Kinghorn, Paul Yun-Farmbrough, Will Varona, Miguel De Llanza Singh, Ryan Buckley, Christopher L. |
| author_facet | Innocenti, Francesco Kinghorn, Paul Yun-Farmbrough, Will Varona, Miguel De Llanza Singh, Ryan Buckley, Christopher L. |
| contents | We introduce JPC, a JAX library for training neural networks with Predictive Coding. JPC provides a simple, fast and flexible interface to train a variety of PC networks (PCNs) including discriminative, generative and hybrid models. Unlike existing libraries, JPC leverages ordinary differential equation solvers to integrate the gradient flow inference dynamics of PCNs. We find that a second-order solver achieves significantly faster runtimes compared to standard Euler integration, with comparable performance on a range of tasks and network depths. JPC also provides some theoretical tools that can be used to study PCNs. We hope that JPC will facilitate future research of PC. The code is available at https://github.com/thebuckleylab/jpc. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2412_03676 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | JPC: Flexible Inference for Predictive Coding Networks in JAX Innocenti, Francesco Kinghorn, Paul Yun-Farmbrough, Will Varona, Miguel De Llanza Singh, Ryan Buckley, Christopher L. Neural and Evolutionary Computing Artificial Intelligence Machine Learning We introduce JPC, a JAX library for training neural networks with Predictive Coding. JPC provides a simple, fast and flexible interface to train a variety of PC networks (PCNs) including discriminative, generative and hybrid models. Unlike existing libraries, JPC leverages ordinary differential equation solvers to integrate the gradient flow inference dynamics of PCNs. We find that a second-order solver achieves significantly faster runtimes compared to standard Euler integration, with comparable performance on a range of tasks and network depths. JPC also provides some theoretical tools that can be used to study PCNs. We hope that JPC will facilitate future research of PC. The code is available at https://github.com/thebuckleylab/jpc. |
| title | JPC: Flexible Inference for Predictive Coding Networks in JAX |
| topic | Neural and Evolutionary Computing Artificial Intelligence Machine Learning |
| url | https://arxiv.org/abs/2412.03676 |