Compositional meta-learning through probabilistic task inference

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Bakermans, Jacob J. W., Tano, Pablo, Riveland, Reidar, Findling, Charles, Pouget, Alexandre
Formato: Preprint
Publicado: 2025
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866916985485393920
author Bakermans, Jacob J. W.
Tano, Pablo
Riveland, Reidar
Findling, Charles
Pouget, Alexandre
author_facet Bakermans, Jacob J. W.
Tano, Pablo
Riveland, Reidar
Findling, Charles
Pouget, Alexandre
contents To solve a new task from minimal experience, it is essential to effectively reuse knowledge from previous tasks, a problem known as meta-learning. Compositional solutions, where common elements of computation are flexibly recombined into new configurations, are particularly well-suited for meta-learning. Here, we propose a compositional meta-learning model that explicitly represents tasks as structured combinations of reusable computations. We achieve this by learning a generative model that captures the underlying components and their statistics shared across a family of tasks. This approach transforms learning a new task into a probabilistic inference problem, which allows for finding solutions without parameter updates through highly constrained hypothesis testing. Our model successfully recovers ground truth components and statistics in rule learning and motor learning tasks. We then demonstrate its ability to quickly infer new solutions from just single examples. Together, our framework joins the expressivity of neural networks with the data-efficiency of probabilistic inference to achieve rapid compositional meta-learning.
format Preprint
id arxiv_https___arxiv_org_abs_2510_01858
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Compositional meta-learning through probabilistic task inference
Bakermans, Jacob J. W.
Tano, Pablo
Riveland, Reidar
Findling, Charles
Pouget, Alexandre
Machine Learning
Neurons and Cognition
To solve a new task from minimal experience, it is essential to effectively reuse knowledge from previous tasks, a problem known as meta-learning. Compositional solutions, where common elements of computation are flexibly recombined into new configurations, are particularly well-suited for meta-learning. Here, we propose a compositional meta-learning model that explicitly represents tasks as structured combinations of reusable computations. We achieve this by learning a generative model that captures the underlying components and their statistics shared across a family of tasks. This approach transforms learning a new task into a probabilistic inference problem, which allows for finding solutions without parameter updates through highly constrained hypothesis testing. Our model successfully recovers ground truth components and statistics in rule learning and motor learning tasks. We then demonstrate its ability to quickly infer new solutions from just single examples. Together, our framework joins the expressivity of neural networks with the data-efficiency of probabilistic inference to achieve rapid compositional meta-learning.
title Compositional meta-learning through probabilistic task inference
topic Machine Learning
Neurons and Cognition
url https://arxiv.org/abs/2510.01858