SNN-Based Online Learning of Concepts and Action Laws in an Open World

TL;DR

This paper introduces a bio-inspired cognitive agent using a spiking neural network for one-shot learning of concepts and action laws, enabling autonomous exploration and adaptation in an open world.

Contribution

It presents a novel SNN-based architecture for one-shot learning of object, situation, and action concepts, including action laws, in an autonomous agent.

Findings

Agent effectively learns and generalizes concepts from minimal data.

Agent rapidly adapts concepts to environmental changes.

Decision-making is based on predicted outcomes from semantic memory.

Abstract

We present the architecture of a fully autonomous, bio-inspired cognitive agent built around a spiking neural network (SNN) implementing the agent's semantic memory. This agent explores its universe and learns concepts of objects/situations and of its own actions in a one-shot manner. While object/situation concepts are unary, action concepts are triples made up of an initial situation, a motor activity, and an outcome. They embody the agent's knowledge of its universe's action laws. Both kinds of concepts have different degrees of generality. To make decisions the agent queries its semantic memory for the expected outcomes of envisaged actions and chooses the action to take on the basis of these predictions. Our experiments show that the agent handles new situations by appealing to previously learned general concepts and rapidly modifies its concepts to adapt to environment changes.