Efficient and robust multi-task learning in the brain with modular latent primitives

TL;DR

This paper introduces a modular recurrent neural network architecture inspired by brain mechanisms, enabling efficient multi-task learning by reusing pre-trained latent modules, resulting in improved robustness and generalization.

Contribution

The authors propose the Modular Latent Primitives (MoLaP) network that leverages pre-trained modules for multi-task learning, inspired by brain-inspired inductive biases, enhancing efficiency and robustness.

Findings

MoLaP enables learning multiple tasks with low parameter updates.

The model shows increased robustness to perturbations.

Generalizes better to new tasks compared to other methods.

Abstract

Biological agents do not have infinite resources to learn new things. For this reason, a central aspect of human learning is the ability to recycle previously acquired knowledge in a way that allows for faster, less resource-intensive acquisition of new skills. In spite of that, how neural networks in the brain leverage existing knowledge to learn new computations is not well understood. In this work, we study this question in artificial recurrent neural networks (RNNs) trained on a corpus of commonly used neuroscience tasks. Combining brain-inspired inductive biases we call functional and structural, we propose a system that learns new tasks by building on top of pre-trained latent dynamics organised into separate recurrent modules. These modules, acting as prior knowledge acquired previously through evolution or development, are pre-trained on the statistics of the full corpus of tasks so as to be independent and maximally informative. The resulting model, we call a Modular Latent Primitives (MoLaP) network, allows for learning multiple tasks while keeping parameter counts, and updates, low. We also show that the skills acquired with our approach are more robust to a broad range of perturbations compared to those acquired with other multi-task learning strategies, and that generalisation to new tasks is facilitated. This work offers a new perspective on achieving efficient multi-task learning in the brain, illustrating the benefits of leveraging pre-trained latent dynamical primitives.