REVIEW 4 cited by
Reward Shaping via Meta-Learning
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
read the original abstract
Reward shaping is one of the most effective methods to tackle the crucial yet challenging problem of credit assignment in Reinforcement Learning (RL). However, designing shaping functions usually requires much expert knowledge and hand-engineering, and the difficulties are further exacerbated given multiple similar tasks to solve. In this paper, we consider reward shaping on a distribution of tasks, and propose a general meta-learning framework to automatically learn the efficient reward shaping on newly sampled tasks, assuming only shared state space but not necessarily action space. We first derive the theoretically optimal reward shaping in terms of credit assignment in model-free RL. We then propose a value-based meta-learning algorithm to extract an effective prior over the optimal reward shaping. The prior can be applied directly to new tasks, or provably adapted to the task-posterior while solving the task within few gradient updates. We demonstrate the effectiveness of our shaping through significantly improved learning efficiency and interpretable visualizations across various settings, including notably a successful transfer from DQN to DDPG.
Forward citations
Cited by 4 Pith papers
-
Learning Upper Lower Value Envelopes to Shape Online RL: A Principled Approach
A two-stage RL framework learns value-function envelopes from offline data and uses them to shape online exploration, yielding regret bounds that improve as offline data grows.
-
Misalignment from Treating Means as Ends
Even a slight mixture of reward and value in a proxy reward can force a reinforcement learning agent to endlessly pursue an instrumental goal, losing all true reward.
-
Meta-Learned Reward Shaping for Reinforcement Learning from Human Feedback
Training a small multi-task reward-shaping network and adding it to the RLHF reward is claimed to improve LLaMA-3-8B alignment across four benchmarks, but the supporting theory is not established.
-
Reward Modeling for Reinforcement Learning-Based LLM Reasoning: Design, Challenges, and Evaluation
A taxonomy-driven survey arguing that reward design is the central mechanism shaping reliable LLM reasoning, with maps of reward paradigms, reward-hacking failure modes, and benchmark pitfalls.
Discussion (0). Sign in to comment.