Learning Temporally-Aware Sample Weights for Preference Optimization (2026.findings-acl)
Copied to clipboard
| Challenge: | Existing methods for preference optimization rely on static functions of instantaneous model states and ignore temporal learning dynamics. |
| Approach: | They propose a framework that meta-learns adaptive weights using three temporal features: reward margin evolution, learning volatility, and reference deviation. |
| Outcome: | The proposed framework achieves statistically significant improvements over baselines on models ranging from 7B to 70B parameters. |
Similar Papers
DORM: Preference Data Weights Optimization for Reward Modeling in LLM Alignment (2025.findings-emnlp)
Copied to clipboard
Rongzhi Zhang, Chenwei Zhang, Xinyang Zhang, Liang Qiu, Haoming Jiang, Yuchen Zhuang, Qingru Zhang, Hyokun Yun, Xian Li, Bing Yin, Tuo Zhao, Chao Zhang
| Challenge: | Existing approaches to align large language models with human preferences are noisy and varying in importance of preference samples. |
| Approach: | a new method enhances reward modeling by learning to dynamically weigh preference data. |
| Outcome: | a new method improves the performance of large language models with human preferences . it initializes data importance and iteratively refines them to maximize validation performance. |
What Makes LLMs Effective Sequential Recommenders? A Study on Preference Intensity and Temporal Context (2026.acl-long)
Copied to clipboard
| Challenge: | Existing preference-alignment approaches rely on binary pairwise comparisons, overlooking preference intensity and temporal context. |
| Approach: | They propose a unified preference optimization framework that maps both explicit and implicit feedback into a common preference signal and constructs adaptive reward margins that jointly account for preference intensity and interaction recency. |
| Outcome: | The proposed framework outperforms state-of-the-art recommendations while maintaining behavioral patterns aligned with human decision-making. |
Beyond Online Sampling: Bridging Offline-to-Online Alignment via Dynamic Data Transformation for LLMs (2025.emnlp-main)
Copied to clipboard
| Challenge: | Direct Preference Optimization (DPO) eliminates complex reward modeling in aligning large language models with human preferences, but its online variant faces significant efficiency bottlenecks due to costly real-time preference sampling and the reward model annotation. |
| Approach: | They propose a framework that transforms static datasets into dynamically adaptive equivalents without the need for an explicit reward model. |
| Outcome: | The proposed approach matches or exceeds the performance of a fully online DPO. |
Dynamic Sampling that Adapts: Self-Aware Iterative Data Persistent Optimization for Mathematical Reasoning (2026.findings-acl)
Copied to clipboard
| Challenge: | Current data selection paradigms rely on static, externally defined metrics, which fail to adapt to the evolving capabilities of models during training. |
| Approach: | They propose a dynamic sampling framework that aligns training data with the model's intrinsic competence by iterating on real-time feedback. |
| Outcome: | Extensive experiments on eight benchmarks show that SAI-DPO outperforms static baselines at most nearly 6 points, achieving state-of-the-art efficiency with significantly less data. |
DynamicFocalPO: Adaptive Focusing Strategy for Preference Optimization (2026.findings-acl)
Copied to clipboard
| Challenge: | Recent preference optimization algorithms such as Direct Preference Optimization (DPO) have become prevalent for aligning large language models with human preferences. |
| Approach: | They propose a preference optimization algorithm that introduces a modulating factor that down-weighs misranked preference pairs and employs focusing strategy that adapts over the course of training. |
| Outcome: | Experiments show that DynamicFocalPO surpasses both DPO and FocalPO on benchmarks including Alpaca Eval 2.0 and Arena-Hard using Mistral-Base-7B and Llama-3-Instruct-8B. |
DISCO Balances the Scales: Adaptive Domain- and Difficulty-Aware Reinforcement Learning on Imbalanced Data (2025.findings-emnlp)
Copied to clipboard
Yuhang Zhou, Jing Zhu, Shengyi Qian, Zhuokai Zhao, Xiyao Wang, Xiaoyu Liu, Ming Li, Paiheng Xu, Wei Ai, Furong Huang
| Challenge: | Large Language Models (LLMs) are increasingly aligned with human preferences through Reinforcement Learning from Human Feedback (RLHF). |
| Approach: | a new study proposes a domain-informed self-consistency policy optimization extension to GRPO that addresses inter-group imbalance. |
| Outcome: | a new extension of GRPO addresses inter-group imbalance with two key innovations . the proposed method outperforms existing GR PO variants by 5% on Qwen3 models . |
MAESTRO: Meta-learning Adaptive Estimation of Scalarization Trade-offs for Reward Optimization (2026.acl-long)
Copied to clipboard
Yang Zhao, Hepeng Wang, Xiao Ding, Yangou Ouyang, Bibo Cai, Kai Xiong, Jinglong Gao, Zhouhao Sun, Li Du, Bing Qin, Ting Liu
| Challenge: | Group-Relative Policy Optimization (GRPO) has emerged as an efficient paradigm for aligning Large Language Models (LLMs), but its efficacy is confined to domains with verifiable ground truths. |
| Approach: | They propose a meta-cognitive orchestration layer that treats reward scalarization as a dynamic latent policy, leveraging the model’s terminal hidden states as 'a semantic bottleneck' . Across seven benchmarks, MAESTRO consistently outperforms single-reward and static multi-objective baselines while preserving the efficiency advantages of GRPO. |
| Outcome: | The proposed model outperforms single-reward and static multi-objective baselines while preserving efficiency advantages. |
Finding the Sweet Spot: Preference Data Construction for Scaling Preference Optimization (2025.acl-long)
Copied to clipboard
| Challenge: | Large language models generate unintended outputs due to their unsupervised nature. |
| Approach: | They propose a method to construct preference pairs of selected and rejected LLMs by repeated random sampling to improve alignment performance. |
| Outcome: | The proposed method improves performance as the sample size increases. |
SeaPO: Strategic Error Amplification for Robust Preference Optimization of Large Language Models (2025.findings-emnlp)
Copied to clipboard
| Challenge: | Existing methods for preference optimization of large language models use pairs of positive and negative samples, but the quality of positive samples may become similar during training, complicating preference learning. |
| Approach: | SeaPO introduces error types commonly occurring in large language models to improve preference learning. |
| Outcome: | SeaPO introduces error types into model Preference Optimization to improve model performance . negative samples are more erroneous than positive samples, and preference-based training mitigates errors . |
Offline Preference Optimization via Maximum Marginal Likelihood Estimation (2026.eacl-long)
Copied to clipboard
| Challenge: | Existing approaches to align Large Language Models with human preferences are complex and unstable. |
| Approach: | They propose a new approach that maximizes the marginal log-likelihood of a preferred text output by using the preference pair as samples for approximation. |
| Outcome: | The proposed approach maximizes the marginal log-likelihood of a preferred text output, using the preference pair as samples for approximation, and forgoes the need for both an explicit reward model and entropy maximization. |