Pith. sign in

REVIEW 7 cited by

Reinforcement Learning with Human Feedback: Learning Dynamic Choices via Pessimism

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

arxiv 2305.18438 v3 pith:LMDUVRXQ submitted 2023-05-29 cs.LG cs.AImath.OCmath.STstat.MLstat.TH

classification cs.LGcs.AImath.OCmath.STstat.MLstat.TH
keywords humanunderlinepolicychoicesdynamicfeedbacklearningmodel
verification ladder T0 review T1 audit T2 compute T3 formal
0 comments
read the original abstract

In this paper, we study offline Reinforcement Learning with Human Feedback (RLHF) where we aim to learn the human's underlying reward and the MDP's optimal policy from a set of trajectories induced by human choices. RLHF is challenging for multiple reasons: large state space but limited human feedback, the bounded rationality of human decisions, and the off-policy distribution shift. In this paper, we focus on the Dynamic Discrete Choice (DDC) model for modeling and understanding human choices. DCC, rooted in econometrics and decision theory, is widely used to model a human decision-making process with forward-looking and bounded rationality. We propose a \underline{D}ynamic-\underline{C}hoice-\underline{P}essimistic-\underline{P}olicy-\underline{O}ptimization (DCPPO) method. \ The method involves a three-stage process: The first step is to estimate the human behavior policy and the state-action value function via maximum likelihood estimation (MLE); the second step recovers the human reward function via minimizing Bellman mean squared error using the learned value functions; the third step is to plug in the learned reward and invoke pessimistic value iteration for finding a near-optimal policy. With only single-policy coverage (i.e., optimal policy) of the dataset, we prove that the suboptimality of DCPPO almost matches the classical pessimistic offline RL algorithm in terms of suboptimality's dependency on distribution shift and dimension. To the best of our knowledge, this paper presents the first theoretical guarantees for off-policy offline RLHF with dynamic discrete choice model.

Discussion (0). Sign in to comment.

Forward citations

Cited by 7 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Sign-SZPO: Provable Preference-based Reinforcement Learning with an Unknown Link Function

    cs.LG 2025-06 conditional novelty 7.0 of 10

    ZSPO provably converges to a stationary policy using only the sign of preference feedback, without knowing the link function between preferences and rewards.

  2. EmoAgent-R1: Towards Multimodal Emotion Understanding with Reinforcement Learning-based Dynamic Agent Specialization

    cs.AI 2026-07 conditional novelty 6.0 of 10

    EmoAgent-R1 combines dynamic agent routing with a token-reweighted GRPO variant (P-GRPO) to reach 77.85% mean on MER-UniBench, exceeding AffectGPT-R1 by 1.90 points.

  3. Multi-Turn On-Policy Distillation with Prefix Replay

    cs.LG 2026-07 conditional novelty 6.0 of 10

    ReOPD offline-distills multi-turn agentic LLMs via teacher-prefix replay plus step-decay sampling, matching online OPD accuracy at ≥4× speed with zero tool calls.

  4. CriticLean: Critic-Guided Reinforcement Learning for Mathematical Formalization

    cs.CL 2025-07 conditional novelty 6.0 of 10

    A critic model trained with reinforcement learning judges semantic correctness of Lean 4 formalizations, and using it as a filter sharply improves autoformalization accuracy.

  5. Thompson Sampling in Online RLHF with General Function Approximation

    cs.LG 2025-05 reject novelty 6.0 of 10

    A model-free posterior sampling algorithm for online RLHF is shown to achieve O(sqrt(T)) regret when the completed function class has low Bellman eluder dimension.

  6. Context Reasoner: Incentivizing Reasoning Capability for Contextualized Privacy and Safety Compliance via Reinforcement Learning

    cs.CL 2025-05 conditional novelty 5.0 of 10

    A Context Reasoner pipeline that cold-starts LLMs on distilled legal reasoning and applies PPO with a rule-based compliance reward improves performance on CI-based legal compliance benchmarks and transfers to general ...

  7. MultiFluxAI Enhancing Platform Engineering with Advanced Agent-Orchestrated Retrieval Systems

    cs.AI 2025-08 reject novelty 4.0 of 10

    The authors claim their MultiFluxAI orchestration framework achieves 95% accuracy and 0-10 ms responses by combining rule-based routing, caching, and graph knowledge stores for multi-service RAG queries.

Pith tools