Quantifying generalization in reinforcement learning
OpenAI Blog
Researchers released CoinRun, a training environment designed to measure how well reinforcement learning agents transfer knowledge to unfamiliar scenarios. The environment sits between simple benchmarks and complex games like Sonic the Hedgehog, offering a calibrated test for generalization capabilities. This tool has resolved previous debates about why some reinforcement learning algorithms struggle with transfer learning despite strong performance on training tasks.
Why it matters
We’re releasing CoinRun, a training environment which provides a metric for an agent’s ability to transfer its experience to novel situations and has already helped clarify a longstanding puzzle in reinforcement learning. CoinRun strikes a desirable balance in complexity: the environment is simpler than traditional platformer games like Sonic the Hedgehog but still poses a worthy generalization challenge for state of the art algorithms.