Predictive Learning Enables Neural Networks to Learn Complex Working Memory Tasks

Thijs Lambik Van der Plas, Sanjay G. Manohar, Tim P. Vogels
Proceedings of The 1st Conference on Lifelong Learning Agents, PMLR 199:518-531, 2022.

Abstract

Brains are thought to engage in predictive learning - learning to predict upcoming stimuli - to construct an internal model of their environment. This is especially notable for spatial navigation, as first described by Tolman’s latent learning tasks. However, predictive learning has also been observed in sensory cortex, in settings unrelated to spatial navigation. Apart from normative frameworks such as active inference or efficient coding, what could be the utility of learning to predict the patterns of occurrence of correlated stimuli? Here we show that prediction, and thereby the construction of an internal model of sequential stimuli, can bootstrap the learning process of a working memory task in a recurrent neural network. We implemented predictive learning alongside working memory match-tasks, and networks emerged to solve the prediction task first by encoding information across time to predict upcoming stimuli, and then eavesdropped on this solution to solve the matching task. Eavesdropping was most beneficial when neural resources were limited. Hence, predictive learning acts as a general neural mechanism to learn to store sensory information that can later be essential for working memory tasks.

Cite this Paper

BibTeX
@InProceedings{pmlr-v199-plas22a, title = {Predictive Learning Enables Neural Networks to Learn Complex Working Memory Tasks}, author = {Van der Plas, Thijs Lambik and Manohar, Sanjay G. and Vogels, Tim P.}, booktitle = {Proceedings of The 1st Conference on Lifelong Learning Agents}, pages = {518--531}, year = {2022}, editor = {Chandar, Sarath and Pascanu, Razvan and Precup, Doina}, volume = {199}, series = {Proceedings of Machine Learning Research}, month = {22--24 Aug}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v199/plas22a/plas22a.pdf}, url = {https://proceedings.mlr.press/v199/plas22a.html}, abstract = {Brains are thought to engage in predictive learning - learning to predict upcoming stimuli - to construct an internal model of their environment. This is especially notable for spatial navigation, as first described by Tolman’s latent learning tasks. However, predictive learning has also been observed in sensory cortex, in settings unrelated to spatial navigation. Apart from normative frameworks such as active inference or efficient coding, what could be the utility of learning to predict the patterns of occurrence of correlated stimuli? Here we show that prediction, and thereby the construction of an internal model of sequential stimuli, can bootstrap the learning process of a working memory task in a recurrent neural network. We implemented predictive learning alongside working memory match-tasks, and networks emerged to solve the prediction task first by encoding information across time to predict upcoming stimuli, and then eavesdropped on this solution to solve the matching task. Eavesdropping was most beneficial when neural resources were limited. Hence, predictive learning acts as a general neural mechanism to learn to store sensory information that can later be essential for working memory tasks.} }
Endnote
%0 Conference Paper %T Predictive Learning Enables Neural Networks to Learn Complex Working Memory Tasks %A Thijs Lambik Van der Plas %A Sanjay G. Manohar %A Tim P. Vogels %B Proceedings of The 1st Conference on Lifelong Learning Agents %C Proceedings of Machine Learning Research %D 2022 %E Sarath Chandar %E Razvan Pascanu %E Doina Precup %F pmlr-v199-plas22a %I PMLR %P 518--531 %U https://proceedings.mlr.press/v199/plas22a.html %V 199 %X Brains are thought to engage in predictive learning - learning to predict upcoming stimuli - to construct an internal model of their environment. This is especially notable for spatial navigation, as first described by Tolman’s latent learning tasks. However, predictive learning has also been observed in sensory cortex, in settings unrelated to spatial navigation. Apart from normative frameworks such as active inference or efficient coding, what could be the utility of learning to predict the patterns of occurrence of correlated stimuli? Here we show that prediction, and thereby the construction of an internal model of sequential stimuli, can bootstrap the learning process of a working memory task in a recurrent neural network. We implemented predictive learning alongside working memory match-tasks, and networks emerged to solve the prediction task first by encoding information across time to predict upcoming stimuli, and then eavesdropped on this solution to solve the matching task. Eavesdropping was most beneficial when neural resources were limited. Hence, predictive learning acts as a general neural mechanism to learn to store sensory information that can later be essential for working memory tasks.
APA
Van der Plas, T.L., Manohar, S.G. & Vogels, T.P.. (2022). Predictive Learning Enables Neural Networks to Learn Complex Working Memory Tasks. Proceedings of The 1st Conference on Lifelong Learning Agents, in Proceedings of Machine Learning Research 199:518-531 Available from https://proceedings.mlr.press/v199/plas22a.html.

Related Material