Learning to Generalize from Sparse and Underspecified Rewards
This addresses the challenge of generalization in weakly-supervised settings like semantic parsing, where rewards are vague, but it is incremental as it builds on existing methods for reward learning.
The paper tackles the problem of learning from sparse and underspecified binary rewards, such as success-failure feedback, by proposing Meta Reward Learning (MeRL) to construct an auxiliary reward function, achieving state-of-the-art results with improvements of 1.2% and 2.4% on WikiTableQuestions and WikiSQL datasets.
We consider the problem of learning from sparse and underspecified rewards, where an agent receives a complex input, such as a natural language instruction, and needs to generate a complex response, such as an action sequence, while only receiving binary success-failure feedback. Such success-failure rewards are often underspecified: they do not distinguish between purposeful and accidental success. Generalization from underspecified rewards hinges on discounting spurious trajectories that attain accidental success, while learning from sparse feedback requires effective exploration. We address exploration by using a mode covering direction of KL divergence to collect a diverse set of successful trajectories, followed by a mode seeking KL divergence to train a robust policy. We propose Meta Reward Learning (MeRL) to construct an auxiliary reward function that provides more refined feedback for learning. The parameters of the auxiliary reward function are optimized with respect to the validation performance of a trained policy. The MeRL approach outperforms our alternative reward learning technique based on Bayesian Optimization, and achieves the state-of-the-art on weakly-supervised semantic parsing. It improves previous work by 1.2% and 2.4% on WikiTableQuestions and WikiSQL datasets respectively.