Unsupervised Hebbian Learning on Point Sets in StarCraft II

TL;DR

This paper introduces an unsupervised Hebbian learning approach for predicting movement of game units in StarCraft II, improving efficiency and accuracy over traditional methods.

Contribution

It proposes a novel Hebbian learning method with neuron activity aware learning and k-Winner-Takes-All for point set feature extraction in RTS games.

Findings

Lower prediction loss compared to self-supervised learning

Significant reduction in computational cost

Effective in modeling unit movement in StarCraft II

Abstract

Learning the evolution of real-time strategy (RTS) game is a challenging problem in artificial intelligent (AI) system. In this paper, we present a novel Hebbian learning method to extract the global feature of point sets in StarCraft II game units, and its application to predict the movement of the points. Our model includes encoder, LSTM, and decoder, and we train the encoder with the unsupervised learning method. We introduce the concept of neuron activity aware learning combined with k-Winner-Takes-All. The optimal value of neuron activity is mathematically derived, and experiments support the effectiveness of the concept over the downstream task. Our Hebbian learning rule benefits the prediction with lower loss compared to self-supervised learning. Also, our model significantly saves the computational cost such as activations and FLOPs compared to a frame-based approach.