# In-Dataset Trajectory Return Regularization for Offline Preference-based Reinforcement Learning

arXiv: https://arxiv.org/abs/2412.09104 · HTML: https://mlsanity.com/paper/2412.09104/

## Key facts

- Cited by 3 papers in the map (data through 14 Oct 2026).
- Placed in Reward modeling and RLHF, one of 10 directions in LLM alignment / RLHF / preference optimisation.
- Builds on 6 of its 49 references that are in the map.
- First posted on arXiv 12 Dec 2024.

## Placed in

- LLM alignment / RLHF / preference optimisation / [Reward modeling and RLHF](https://mlsanity.com/alignment/reward-modeling-and-rlhf/)
- Offline RL and imitation learning from data / [Offline RL algorithms](https://mlsanity.com/offlinerl/offline-rl-algorithms/)

## Builds on (most cited first, 6 of 6)

- [OpenAI Gym](https://mlsanity.com/paper/1606.01540/) (2016)
- [D4RL: Datasets for Deep Data-Driven Reinforcement Learning](https://mlsanity.com/paper/2004.07219/) (2020)
- [PEBBLE: Feedback-Efficient Interactive Reinforcement Learning via Relabeling Experience and Unsupervised Pre-training](https://mlsanity.com/paper/2106.05091/) (2021)
- [Is DPO Superior to PPO for LLM Alignment? A Comprehensive Study](https://mlsanity.com/paper/2404.10719/) (2024)
- [Return-Aligned Decision Transformer](https://mlsanity.com/paper/2402.03923/) (2024)
- [Hindsight Preference Learning for Offline Preference-based Reinforcement Learning](https://mlsanity.com/paper/2407.04451/) (2024)

## Built on by (most cited first, 3 of 3)

- [From Reward-Free Representations to Preferences: Rethinking Offline Preference-Based Reinforcement Learning](https://mlsanity.com/paper/2606.01123/) (2026)
- [Diffusion Classifier-Driven Reward for Offline Preference-based Reinforcement Learning](https://mlsanity.com/paper/2503.01143/) (2025)
- [Conditional Sequence Modeling for Safe Reinforcement Learning](https://mlsanity.com/paper/2602.08584/) (2026)

## Abstract (arXiv)

Offline preference-based reinforcement learning (PbRL) typically operates in two phases: first, use human preferences to learn a reward model and annotate rewards for a reward-free offline dataset; second, learn a policy by optimizing the learned reward via offline RL. However, accurately modeling step-wise rewards from trajectory-level preference feedback presents inherent challenges. The reward bias introduced, particularly the overestimation of predicted rewards, leads to optimistic trajectory stitching, which undermines the pessimism mechanism critical to the offline RL phase. To address this challenge, we propose In-Dataset Trajectory Return Regularization (DTR) for offline PbRL, which leverages conditional sequence modeling to mitigate the risk of learning inaccurate trajectory stitching under reward bias. Specifically, DTR employs Decision Transformer and TD-Learning to strike a balance between maintaining fidelity to the behavior policy with high in-dataset trajectory returns and selecting optimal actions based on high reward labels. Additionally, we introduce an ensemble normalization technique that effectively integrates multiple reward models, balancing the tradeoff between reward differentiation and accuracy. Empirical evaluations on various benchmarks demonstrate the superiority of DTR over other state-of-the-art baselines.