Meta-Dynamical State Space Models for Integrative Neural Data Analysis
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_ | 1866912313907347456 |
|---|---|
| author | Vermani, Ayesha Nassar, Josue Jeon, Hyungju Dowling, Matthew Park, Il Memming |
| author_facet | Vermani, Ayesha Nassar, Josue Jeon, Hyungju Dowling, Matthew Park, Il Memming |
| contents | Learning shared structure across environments facilitates rapid learning and adaptive behavior in neural systems. This has been widely demonstrated and applied in machine learning to train models that are capable of generalizing to novel settings. However, there has been limited work exploiting the shared structure in neural activity during similar tasks for learning latent dynamics from neural recordings. Existing approaches are designed to infer dynamics from a single dataset and cannot be readily adapted to account for statistical heterogeneities across recordings. In this work, we hypothesize that similar tasks admit a corresponding family of related solutions and propose a novel approach for meta-learning this solution space from task-related neural activity of trained animals. Specifically, we capture the variabilities across recordings on a low-dimensional manifold which concisely parametrizes this family of dynamics, thereby facilitating rapid learning of latent dynamics given new recordings. We demonstrate the efficacy of our approach on few-shot reconstruction and forecasting of synthetic dynamical systems, and neural recordings from the motor cortex during different arm reaching tasks. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2410_05454 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Meta-Dynamical State Space Models for Integrative Neural Data Analysis Vermani, Ayesha Nassar, Josue Jeon, Hyungju Dowling, Matthew Park, Il Memming Machine Learning Neurons and Cognition Learning shared structure across environments facilitates rapid learning and adaptive behavior in neural systems. This has been widely demonstrated and applied in machine learning to train models that are capable of generalizing to novel settings. However, there has been limited work exploiting the shared structure in neural activity during similar tasks for learning latent dynamics from neural recordings. Existing approaches are designed to infer dynamics from a single dataset and cannot be readily adapted to account for statistical heterogeneities across recordings. In this work, we hypothesize that similar tasks admit a corresponding family of related solutions and propose a novel approach for meta-learning this solution space from task-related neural activity of trained animals. Specifically, we capture the variabilities across recordings on a low-dimensional manifold which concisely parametrizes this family of dynamics, thereby facilitating rapid learning of latent dynamics given new recordings. We demonstrate the efficacy of our approach on few-shot reconstruction and forecasting of synthetic dynamical systems, and neural recordings from the motor cortex during different arm reaching tasks. |
| title | Meta-Dynamical State Space Models for Integrative Neural Data Analysis |
| topic | Machine Learning Neurons and Cognition |
| url | https://arxiv.org/abs/2410.05454 |