Alignment through Meta-Weighted Online Sampling: Bridging the Gap between Data Generation and Preference Optimization

• URL: http://arxiv.org/abs/2509.23371v1
• Date: Sat, 27 Sep 2025 15:38:24 GMT
• Title: Alignment through Meta-Weighted Online Sampling: Bridging the Gap between Data Generation and Preference Optimization
• Abstract summary: Preference optimization is crucial for aligning large language models with human values and intentions.<n>Existing methods attempt to reduce this gap using statics or decoupled online sampling strategies.<n>We propose Meta-Weighted Adaptive Preference Optimization (MetaAPO), a novel framework that dynamically couples data generation with model training.
• Score: 32.84452172765487
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Preference optimization is crucial for aligning large language models (LLMs) with human values and intentions. A significant challenge in this process is the distribution mismatch between pre-collected offline preference data and the evolving model policy. Existing methods attempt to reduce this gap using static heuristics or decoupled online sampling strategies, but they often fail to adapt to the model's dynamic learning state. To bridge this gap, we propose Meta-Weighted Adaptive Preference Optimization (MetaAPO), a novel framework that dynamically couples data generation with model training. MetaAPO employs a lightweight meta-learner, as an "alignment gap estimator", to evaluate the potential benefits of on-policy sampling in relation to offline data. This guides targeted online generation and assigns sample-wise meta-weights to the optimization objective, dynamically balancing the quality and distribution of online and offline data. Experiments on AlpacaEval 2, Arena-Hard and MT-Bench demonstrate that MetaAPO consistently outperforms existing preference optimization approaches across various settings, while reducing 42% in online annotation costs.

Related papers

• Metadata-Free Meta-Reweighted Direct Preference Optimization under Noisy Preference Labels [4.826561433595647]
  We propose a bilevel optimization framework and prove, under certain assumptions, that it can recover the DPO optimum under clean data.<n>Considering that high-quality metadata may be difficult to obtain, we propose a task-agnostic meta-knowledge-driven method.<n> Experiments on TL;DR summarization and Anthropic HH single-turn dialogue show that the proposed method improves training performance over multiple DPO baselines under different noise rates.
  arXiv  Detail & Related papers  (2026-07-09T09:20:25Z)
• Joint Optimization of Training and Inference in Federated Edge Learning via Constrained Multi-Objective Deep Reinforcement Learning [9.37821062000313]
  Federated edge learning (FEEL) has emerged as a promising paradigm for achieving edge intelligence (EI)<n>We put forth an online optimization framework that jointly manages federated training and inference on resource-constrained edge devices.
  arXiv  Detail & Related papers  (2026-05-25T14:51:07Z)
• Two-Stage Optimizer-Aware Online Data Selection for Large Language Models [49.576993784867035]
  We propose a principled-aware framework for gradient-based online data selection and reweighting in large language model fine-tuning.<n>Our key idea is to view online selection not as static sample ranking, but as shaping the next target-oriented update under the state.<n> Experiments show that our method consistently improves convergence and downstream performance over existing online data selection baselines under the same data budget.
  arXiv  Detail & Related papers  (2026-03-08T21:46:16Z)
• Adaptive Sample Scheduling for Direct Preference Optimization [37.54165341391688]
  We introduce a novel problem: Sample Scheduling for DPO.<n>It aims to dynamically and adaptively schedule training samples based on the model's evolving states.<n>We propose SamS, an efficient and effective algorithm that adaptively selects samples in each training batch.
  arXiv  Detail & Related papers  (2025-06-08T10:26:09Z)
• Leveraging Robust Optimization for LLM Alignment under Distribution Shifts [52.983390470606146]
  Preference alignment methods are increasingly critical for steering large language models to generate outputs consistent with human values.<n>We propose a novel distribution-aware optimization framework that improves preference alignment despite such shifts.
  arXiv  Detail & Related papers  (2025-04-08T09:14:38Z)
• Model-Based Offline Reinforcement Learning with Adversarial Data Augmentation [36.9134885948595]
  We introduce Model-based Offline Reinforcement learning with AdversariaL data augmentation.<n>In MORAL, we replace the fixed horizon rollout by employing adversaria data augmentation to execute alternating sampling with ensemble models.<n>Experiments on D4RL benchmark demonstrate that MORAL outperforms other model-based offline RL methods in terms of policy learning and sample efficiency.
  arXiv  Detail & Related papers  (2025-03-26T07:24:34Z)
• Dynamic Noise Preference Optimization for LLM Self-Improvement via Synthetic Data [51.62162460809116]
  We introduce Dynamic Noise Preference Optimization (DNPO) to ensure consistent improvements across iterations.<n>In experiments with Zephyr-7B, DNPO consistently outperforms existing methods, showing an average performance boost of 2.6%.<n> DNPO shows a significant improvement in model-generated data quality, with a 29.4% win-loss rate gap compared to the baseline in GPT-4 evaluations.
  arXiv  Detail & Related papers  (2025-02-08T01:20:09Z)
• SoPo: Text-to-Motion Generation Using Semi-Online Preference Optimization [49.76016928551215]
  We focus on fine-tuning text-to-motion models to consistently favor high-quality, human-preferred motions.<n>We introduce Semi-online Preference Optimization (SoPo), a DPO-based method for training text-to-motion models.<n>Experiments demonstrate that SoPo outperforms other preference alignment methods.
  arXiv  Detail & Related papers  (2024-12-06T14:50:38Z)
• Margin Matching Preference Optimization: Enhanced Model Alignment with Granular Feedback [64.67540769692074]
  Large language models (LLMs) fine-tuned with alignment techniques, such as reinforcement learning from human feedback, have been instrumental in developing some of the most capable AI systems to date.<n>We introduce an approach called Margin Matching Preference Optimization (MMPO), which incorporates relative quality margins into optimization, leading to improved LLM policies and reward models.<n>Experiments with both human and AI feedback data demonstrate that MMPO consistently outperforms baseline methods, often by a substantial margin, on popular benchmarks including MT-bench and RewardBench.
  arXiv  Detail & Related papers  (2024-10-04T04:56:11Z)
• WarpAdam: A new Adam optimizer based on Meta-Learning approach [0.0]
  This study introduces an innovative approach that merges the 'warped gradient descend' concept from Meta Learning with the Adam. By introducing a learnable distortion matrix P within the adaptation matrix P, we aim to enhance the model's capability across diverse data distributions. Our research showcases potential of this novel approach through theoretical insights and empirical evaluations.
  arXiv  Detail & Related papers  (2024-09-06T12:51:10Z)
• Bridging and Modeling Correlations in Pairwise Data for Direct Preference Optimization [75.1240295759264]
  We propose an effective framework for Bridging and Modeling Correlations in pairwise data, named BMC.<n>We increase the consistency and informativeness of the pairwise preference signals through targeted modifications.<n>We identify that DPO alone is insufficient to model these correlations and capture nuanced variations.
  arXiv  Detail & Related papers  (2024-08-14T11:29:47Z)
• WPO: Enhancing RLHF with Weighted Preference Optimization [40.07940023654452]
  Reinforcement learning from human feedback (RLHF) is a promising solution to align large language models (LLMs) more closely with human values. Off-policy preference optimization often suffers from a distributional gap between the policy used for data collection and the target policy, leading to suboptimal optimization. We propose a novel strategy to mitigate this problem by simulating on-policy learning with off-policy preference data.
  arXiv  Detail & Related papers  (2024-06-17T17:59:13Z)
• Self-Augmented Preference Optimization: Off-Policy Paradigms for Language Model Alignment [104.18002641195442]
  We introduce Self-Augmented Preference Optimization (SAPO), an effective and scalable training paradigm that does not require existing paired data. Building on the self-play concept, which autonomously generates negative responses, we further incorporate an off-policy learning pipeline to enhance data exploration and exploitation.
  arXiv  Detail & Related papers  (2024-05-31T14:21:04Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.