Width-based Lookaheads with Learnt Base Policies and Heuristics Over the Atari-2600 Benchmark
This work addresses performance bottlenecks in reinforcement learning for complex video games, offering incremental improvements over existing width-based methods.
The authors tackled the problem of improving width-based planning and learning algorithms for Atari-2600 games, resulting in a new algorithm, N-CPL, that outperforms previous methods like π-IW(1), π-IW(1)+, and π-HIW(n, 1), particularly in games with large branching factors and sparse rewards.
We propose new width-based planning and learning algorithms inspired from a careful analysis of the design decisions made by previous width-based planners. The algorithms are applied over the Atari-2600 games and our best performing algorithm, Novelty guided Critical Path Learning (N-CPL), outperforms the previously introduced width-based planning and learning algorithms $π$-IW(1), $π$-IW(1)+ and $π$-HIW(n, 1). Furthermore, we present a taxonomy of the Atari-2600 games according to some of their defining characteristics. This analysis of the games provides further insight into the behaviour and performance of the algorithms introduced. Namely, for games with large branching factors, and games with sparse meaningful rewards, N-CPL outperforms $π$-IW, $π$-IW(1)+ and $π$-HIW(n, 1).