Should I Stay or Should I Go: Coordinating Biological Needs with Continuously-updated Assessments of the Environment

TL;DR

This paper introduces Wanderer, a model of autonomous systems that adaptively coordinate internal needs with environmental assessments, balancing exploration and caution based on internal states and learned associations.

Contribution

The paper presents Wanderer, a novel model integrating biological needs, environmental assessment, and associative learning to guide adaptive exploration behavior.

Findings

Wanderer adjusts exploration based on food and predator presence.

It forms associations between neutral stimuli and salient events.

The model demonstrates adaptive short-term and long-term behavior.

Abstract

This paper presents Wanderer, a model of how autonomous adaptive systems coordinate internal biological needs with moment-by-moment assessments of the probabilities of events in the external world. The extent to which Wanderer moves about or explores its environment reflects the relative activations of two competing motivational sub-systems: one represents the need to acquire energy and it excites exploration, and the other represents the need to avoid predators and it inhibits exploration. The environment contains food, predators, and neutral stimuli. Wanderer responds to these events in a way that is adaptive in the short turn, and reassesses the probabilities of these events so that it can modify its long term behaviour appropriately. When food appears, Wanderer be-comes satiated and exploration temporarily decreases. When a predator appears, Wanderer both decreases exploration in the short term, and becomes more "cautious" about exploring in the future. Wanderer also forms associations between neutral features and salient ones (food and predators) when they are present at the same time, and uses these associations to guide its behaviour.