full-FORCE: A Target-Based Method for Training Recurrent Networks

TL;DR

full-FORCE is a novel target-based training method for recurrent networks that enhances task performance, robustness, and flexibility by leveraging a secondary network to generate suitable target dynamics.

Contribution

It introduces a second network during training to generate target dynamics, enabling more efficient and robust training of recurrent networks for complex tasks.

Findings

Networks trained with full-FORCE perform with fewer neurons.

Full-FORCE networks exhibit greater noise robustness.

Adding input signals as hints extends task learning capabilities.

Abstract

Trained recurrent networks are powerful tools for modeling dynamic neural computations. We present a target-based method for modifying the full connectivity matrix of a recurrent network to train it to perform tasks involving temporally complex input/output transformations. The method introduces a second network during training to provide suitable "target" dynamics useful for performing the task. Because it exploits the full recurrent connectivity, the method produces networks that perform tasks with fewer neurons and greater noise robustness than traditional least-squares (FORCE) approaches. In addition, we show how introducing additional input signals into the target-generating network, which act as task hints, greatly extends the range of tasks that can be learned and provides control over the complexity and nature of the dynamics of the trained, task-performing network.