Executive Function: A Contrastive Value Policy for Resampling and Relabeling Perceptions via Hindsight Summarization?

TL;DR

This paper proposes a novel model of executive function as a contrastive value policy that resamples and relabels perceptions through hindsight summarization, aiming to improve few-shot continual learning and explain human neurocognitive processes.

Contribution

It introduces a new theoretical framework linking executive function to a contrastive policy with hindsight-based perception relabeling, grounded in evolutionary principles.

Findings

Model explains human few-shot learning behaviors

Demonstrates improved continual learning performance

Provides insights into neuroanatomical observations

Abstract

We develop the few-shot continual learning task from first principles and hypothesize an evolutionary motivation and mechanism of action for executive function as a contrastive value policy which resamples and relabels perception data via hindsight summarization to minimize attended prediction error, similar to an online prompt engineering problem. This is made feasible by the use of a memory policy and a pretrained network with inductive biases for a grammar of learning and is trained to maximize evolutionary survival. We show how this model of executive function can be used to implement hypothesis testing as a stream of consciousness and may explain observations of human few-shot learning and neuroanatomy.