PUZZLES: A Benchmark for Neural Algorithmic Reasoning
This work addresses the need for benchmarks to foster progress in algorithmic reasoning for reinforcement learning researchers, though it is incremental as it builds on existing puzzle collections.
The authors introduced PUZZLES, a benchmark based on Simon Tatham's Portable Puzzle Collection with 40 diverse logic puzzles, to advance algorithmic and logical reasoning in reinforcement learning, and they evaluated various RL algorithms to provide baseline comparisons.
Algorithmic reasoning is a fundamental cognitive ability that plays a pivotal role in problem-solving and decision-making processes. Reinforcement Learning (RL) has demonstrated remarkable proficiency in tasks such as motor control, handling perceptual input, and managing stochastic environments. These advancements have been enabled in part by the availability of benchmarks. In this work we introduce PUZZLES, a benchmark based on Simon Tatham's Portable Puzzle Collection, aimed at fostering progress in algorithmic and logical reasoning in RL. PUZZLES contains 40 diverse logic puzzles of adjustable sizes and varying levels of complexity; many puzzles also feature a diverse set of additional configuration parameters. The 40 puzzles provide detailed information on the strengths and generalization capabilities of RL agents. Furthermore, we evaluate various RL algorithms on PUZZLES, providing baseline comparisons and demonstrating the potential for future research. All the software, including the environment, is available at https://github.com/ETH-DISCO/rlp.