English

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

Neural and Evolutionary Computing 2018-07-04 v1 Machine Learning Neurons and Cognition Machine Learning

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.

Keywords

Cite

@article{arxiv.1710.03070,
  title  = {full-FORCE: A Target-Based Method for Training Recurrent Networks},
  author = {Brian DePasquale and Christopher J. Cueva and Kanaka Rajan and G. Sean Escola and L. F. Abbott},
  journal= {arXiv preprint arXiv:1710.03070},
  year   = {2018}
}

Comments

20 pages, 8 figures

R2 v1 2026-06-22T22:07:32.299Z