A controller-peripheral architecture and costly energy principle for learning

Abstract

Complex behavior is supported by the coordination of multiple brain regions. We propose coordination is achieved by a controller-peripheral architecture in which peripherals (e.g., the ventral visual stream) aim to supply needed inputs to their controllers (e.g., the hippocampus and prefrontal cortex) while expending minimal resources. We developed a formal model within this framework to address how multiple brain regions coordinate to support rapid learning from a few example images. The model captured how higher-level activity in the controller shaped lower-level visual representations, affecting their precision and sparsity in a manner that paralleled brain measures. Alternative models optimized by gradient descent irrespective of architectural constraints could not account for human behavior or brain responses, and, typical of standard deep learning approaches, were unstable trial-by-trial learners.

Effects of noise and metabolic cost on cortical task representations

“…Several factors could have a strong impact on whether irrelevant stimuli affect task performance. For example, levels of neural noise in the circuit as well as energy constraints and the metabolic costs of overall neural activity 15,[18][19][20][21][22][23][24][25][26] can affect how stimuli are represented in a neural circuit. Indeed, both noise and metabolic costs are factors that biological circuits must contend with [27][28][29][30] .…”

Adaptive stretching of representations across brain regions and deep learning model layers

“…While models of goal-directed attention contain dedicated control systems that selectively weight relevant dimensions [7,9,8,19,20], one possibility is that these control systems themselves can be learned by nonhuman animals and ANNs, enabling them to reconfigure themselves in response to task cues. According to this hypothesis, brains and ANNs are powerful statistical learning machines that build control structures that adaptively stretch learnt representations along relevant dimensions to facilitate task performance.…”