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

2024-11-19

AI Summary

To address the challenges of one-shot learning, dynamic generalization, and continual adaptation for Spiking Neural Networks (SNNs) in open-world environments, this paper proposes a semantic-memory-based autonomous cognitive agent framework. The method models behavioral regularities as “initial-state–action–outcome” semantic triples and embeds them into an online semantic memory system within the SNN. It integrates one-shot concept abstraction, hierarchical generalization, and prediction-driven action selection to enable rapid knowledge reuse in novel situations and real-time concept revision—within a few steps—upon environmental changes. Experiments demonstrate that the agent acquires object/context concepts and behavioral rules from a single observation, achieves efficient cross-environment transfer, and autonomously updates its internal model under distributional shifts. This significantly enhances the cognitive flexibility and robustness of SNNs in open-world continual learning scenarios.

Technology Category

• Cognitive Modeling & Cognitive Systems: Neural Spike Coding
• Search and Optimization: Learning to Search
• Machine Learning: Neuro-Symbolic Learning

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Agentic search
• Responsible Web: Machine-in-the-loop, human agency and autonomy

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. The 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 actions 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.

Problem

• Spiking Neural Networks
• Lifelong Learning
• Adaptability in Unknown Environments

Innovation

• Spiking Neural Network
• Autonomous Learning
• Adaptive Decision-Making