Boosting expert ensembles for rapid concept recall

摘要

Many learning tasks in adversarial domains tend to be highly dependent on the opponent. Predefined strate-gies optimized for play against a specific opponent are not likely to succeed when employed against another opponent. Learning a strategy for each new opponent from scratch, though, is inefficient as one is likely to encounter the same or similar opponents again. We call this particular variant of inductive transfer a con-cept recall problem. We present an extension to Ad-aBoost called ExpBoost that is especially designed for such a sequential learning tasks. It automatically bal-ances between an ensemble of experts each trained on one known opponent and learning the concept of the new opponent. We present and compare results of Exp-Boost and other algorithms on both synthetic data and in a simulated robot soccer task. ExpBoost can rapidly adjust to new concepts and achieve performance com-parable to a classifier trained exclusively on a particular opponent with far more data.