paper-with-me

홈 › Papers

Objective Mismatch in Model-based Reinforcement Learning

2020-02-11 · ICLR 2020 1 · Nathan Lambert, Brandon Amos, Omry Yadan, Roberto Calandra

Model-based reinforcement learning (MBRL) has been shown to be a powerful framework for data-efficiently learning control of continuous tasks. Recent work in MBRL has mostly focused on using more advanced function approximators and planning schemes, with little development of the general framework. In this paper, we identify a fundamental issue of the standard MBRL framework -- what we call the objective mismatch issue. Objective mismatch arises when one objective is optimized in the hope that a second, often uncorrelated, metric will also be optimized. In the context of MBRL, we characterize the objective mismatch between training the forward dynamics model w.r.t.~the likelihood of the one-step ahead prediction, and the overall goal of improving performance on a downstream control task. For example, this issue can emerge with the realization that dynamics models effective for a specific task do not necessarily need to be globally accurate, and vice versa globally accurate models might not be sufficiently accurate locally to obtain good control performance on a specific task. In our experiments, we study this objective mismatch issue and demonstrate that the likelihood of one-step ahead predictions is not always correlated with control performance. This observation highlights a critical limitation in the MBRL framework which will require further research to be fully understood and addressed. We propose an initial method to mitigate the mismatch issue by re-weighting dynamics model training. Building on it, we conclude with a discussion about other potential directions of research for addressing this issue.

📄 PDF Abstract BibTeX arXiv:2002.04523

Code (2)

evgenii-nikishin/omd jax
ran-weii/objective_mismatch_papers

Tasks

modelModel-based Reinforcement Learningreinforcement-learningReinforcement LearningReinforcement Learning (RL)

Similar Papers 제목 키워드 기반

A Unified View on Solving Objective Mismatch in Model-Based Reinforcement Learning

2023-10-10 · Ran Wei, Nathan Lambert, Anthony McDonald, Alfredo Garcia 외

Model-based Reinforcement Learning (MBRL) aims to make agents more sample-efficient, adaptive, and explainable by learning an explicit model of the environment. While the capabilities of MBRL agents have significantly im…

Model-based Reinforcement Learning

The Alignment Ceiling: Objective Mismatch in Reinforcement Learning from Human Feedback

2023-10-31 · Nathan Lambert, Roberto Calandra

Reinforcement learning from human feedback (RLHF) has emerged as a powerful technique to make large language models (LLMs) more capable in complex settings. RLHF proceeds as collecting human preference data, training a r…

GSM8KMMLUModel-based Reinforcement Learningreinforcement-learning

BECAUSE: Bilinear Causal Representation for Generalizable Offline Model-based Reinforcement Learning

2024-07-15 · Haohong Lin, Wenhao Ding, Jian Chen, Laixi Shi 외

Offline model-based reinforcement learning (MBRL) enhances data efficiency by utilizing pre-collected datasets to learn models and policies, especially in scenarios where exploration is costly or infeasible. Nevertheless…

Model-based Reinforcement LearningOffline RL

The Mirage of Optimizing Training Policies: Monotonic Inference Policies as the Real Objective for LLM Reinforcement Learning

2026-06-28 · Jing Liang, Hongyao Tang, Yi Ma, Yancheng He 외 hf

Reinforcement learning (RL) has gained growing attention in large language model (LLM) post-training, yet RL training remains fragile and can suffer from instability or collapse. One vital cause is training-inference mis…

Reinforcement Learning

Maximum Likelihood Reinforcement Learning

2026-02-02 · Fahim Tajwar, Guanning Zeng, Yueer Zhou, Yuda Song 외 arxiv

Reinforcement learning (RL) is the method of choice for training models in setups where the objective function can only be evaluated by sampling from the model. Our key observation is that when the feedback is terminal a…

Reinforcement LearningCode Generation