| Challenge: | Prior work on role-playing agents relies on supervised fine-tuning or reinforcement learning with scalarized rewards, but these approaches do not address the coordination of multiple reward dimensions during optimization. |
| Approach: | They propose a reinforcement-learning framework that enables multi-dimensional, fine-grained rubric optimization for general RPAs. |
| Outcome: | Experiments on PersonaGym and RoleMRC show that MOA improves multi-dimensional role-playing performance over supervised and standard RL baselines. |
Similar Papers
Beyond One-Preference-Fits-All Alignment: Multi-Objective Direct Preference Optimization (2024.findings-acl)
Copied to clipboard
| Challenge: | Recent approaches to language model alignment assume homogeneous human preferences, but actual human preferences vary widely and are hard to satisfy with a single language model. |
| Approach: | They propose an RL-free extension of Direct Preference Optimization (DPO) that folds language modeling directly into reward modeling and trains language models as collective reward models that combine all objectives with specific weights. |
| Outcome: | The proposed method matches or outperforms existing methods in safety alignment and long-form question answering. |
Self-Improvement Towards Pareto Optimality: Mitigating Preference Conflicts in Multi-Objective Alignment (2025.findings-acl)
Copied to clipboard
| Challenge: | Existing approaches to optimize large language models with human preferences suffer from preference conflicts in the data. |
| Approach: | They propose to construct Pareto-optimal responses to resolve preference conflicts by using a self-improving DPO framework that enables LLMs to self-generate and select Paret-optimized responses. |
| Outcome: | The proposed framework achieves superior Pareto Front performance over baselines on two datasets. |
Deep Reinforcement Learning for Entity Alignment (2022.findings-acl)
Copied to clipboard
| Challenge: | Entity alignment (EA) methods identify the aligned entities based on cosine similarity, ignoring the semantics underlying the embeddings themselves. |
| Approach: | They propose to model entity alignment as a sequential decision-making task where an agent sequentially decides whether two entities are matched or mismatched based on representation vectors. |
| Outcome: | The proposed framework consistently advances the performance of several state-of-the-art methods, with a maximum improvement of 31.1% on Hits@1. |
RoleMRC: A Fine-Grained Composite Benchmark for Role-Playing and Instruction-Following (2025.findings-acl)
Copied to clipboard
| Challenge: | Existing role-playing datasets mostly contribute to controlling role style and knowledge boundaries, but overlook role-following in instruction-follower scenarios. |
| Approach: | They propose a fine-grained role-playing and instruction-following composite benchmark, named RoleMRC, which includes multi-turn dialogues between ideal roles and humans, including free chats or discussions upon given passages . |
| Outcome: | The proposed model improves instruction-following without compromising general role-playing and reasoning capabilities. |
Beyond ’Aha!’: Toward Systematic Meta-Abilities Alignment in Large Reasoning Models (2026.findings-acl)
Copied to clipboard
| Challenge: | Prior work has shown that outcome-based reinforcement learning (RL) can incidentally elicit advanced reasoning behaviors such as self-correction, backtracking, and verification. |
| Approach: | They explicitly align large reasoning models with three meta-abilities: deduction, induction, and abduction, using automatically generated, self-verifiable tasks. |
| Outcome: | The proposed model aligns models with deduction, induction, and abduction meta-abilities using automatically generated, self-verifiable tasks. |
Gradient-Adaptive Policy Optimization: Towards Multi-Objective Alignment of Large Language Models (2025.acl-long)
Copied to clipboard
| Challenge: | Reinforcement learning from human feedback (RLHF) has emerged as a powerful technique for aligning large language models (LLMs) with human preferences. |
| Approach: | They propose a novel algorithm that uses multiple-gradient descent to optimize LLMs with diverse preferences to maximize trade-offs between objectives. |
| Outcome: | The proposed approach incorporates user preferences across different objectives and achieves Pareto solutions that better align with the user’s specific needs. |
Arithmetic Control of LLMs for Diverse User Preferences: Directional Preference Alignment with Multi-Objective Rewards (2024.acl-long)
Copied to clipboard
| Challenge: | Reinforcement Learning from Human Feedback (RLHF) relies on scalar rewards to capture user preferences. |
| Approach: | They propose a framework that integrates multi-objective reward modeling to represent diverse preference profiles. |
| Outcome: | The proposed method improves performance across reward objectives and targets. |
Fine-Mem: Fine-Grained Feedback Alignment for Long-Horizon Memory Management (2026.acl-long)
Copied to clipboard
Weitao Ma, Xiaocheng Feng, Lei Huang, Xiachong Feng, Zhanyu Ma, Jun Xu, Jiuchong Gao, Jinghua Hao, Renqing He, Bing Qin
| Challenge: | Existing approaches to memory management rely on final task performance as the primary reward, resulting in severe reward sparsity and ineffective credit assignment. |
| Approach: | They propose a framework for fine-grained feedback alignment using a Chunk-level step reward and Evidence-Anchored Reward Attribution to redistribute global rewards based on memory items utilized as evidence in reasoning. |
| Outcome: | The proposed framework outperforms baselines and supports generalization across different model configurations and backbones. |
Reinforcement Learning with Supervised Alignment (2025.findings-emnlp)
Copied to clipboard
| Challenge: | Supervised fine-tuning (SFT) is a widely used method for adapting Large Language Models to specific tasks. |
| Approach: | They propose a method that uses supervised fine-tuning to train a reward model for reinforcement learning. |
| Outcome: | The proposed method outperforms existing methods on in-domain benchmarks but surpasses them 50 times on out-of-domain and cross-task evaluations. |
RoleCDE: Benchmarking and Mitigating Role–Alignment Trade-offs in Role-Playing Agents (2026.findings-acl)
Copied to clipboard
| Challenge: | Existing benchmarks for role-playing agents only evaluate surface-level fidelity and provide limited insight into decision making under role–alignment value conflicts. |
| Approach: | They propose a benchmark to evaluate RPAs under role–alignment value conflicts . they use 8k diverse role profiles and 240k dilemma instances to evaluate role-aware decision making . |
| Outcome: | The proposed benchmark covers 8k diverse role profiles and scenarios and nearly 240k dilemma instances across three difficulty levels and eight role categories. |