Multi-tasking via baseline control in recurrent neural networks

TL;DR

This paper demonstrates that baseline modulations in recurrent neural networks can induce new dynamic phases and enable multi-tasking without additional training, offering insights into brain function and AI.

Contribution

It introduces a reservoir computing approach showing how baseline modulations facilitate multi-tasking and dynamic phase changes in recurrent neural networks without coupling optimization.

Findings

Baseline modulations induce chaos enhancement and hysteresis.

Reservoir networks perform multiple tasks without optimization.

Baseline control influences cortical activity and AI design.

Abstract

Changes in an animal's behavioral state, such as arousal and movements, induce {complex modulations of the baseline input currents to sensory areas, eliciting sensory modality-specific effects. A simple computational principle explaining the effects of baseline modulations to recurrent cortical circuits is lacking. We investigate the benefits of baseline modulations using a reservoir computing approach in recurrent neural networks with random couplings. Baseline modulations unlock a set of new network phases and phenomena, including chaos enhancement, neural hysteresis and ergodicity breaking. Strikingly, baseline modulations enable reservoir networks to perform multiple tasks, without any optimization of the network couplings.} Baseline control of network dynamics opens new directions for brain-inspired artificial intelligence and sheds new light on behavioral modulations of cortical activity.