Papers by Yijiang River Dong
When Personalization Meets Reality: A Multi-Faceted Analysis of Personalized Preference Learning (2025.findings-emnlp)
Copied to clipboard
| Challenge: | Reinforcement Learning from Human Feedback assumes homogeneous preferences across users . personalization can introduce up to 20% safety misalignment . |
| Approach: | They propose a framework to assess personalized preference learning by tailoring preferences for users . they compare eight personalization methods across three preference datasets . |
| Outcome: | The proposed framework measures performance, fairness, unintended effects, adaptability across preferences . performance differences between personalization methods could reach 36% when users strongly disagree . |
UNDIAL: Self-Distillation with Adjusted Logits for Robust Unlearning in Large Language Models (2025.naacl-long)
Copied to clipboard
| Challenge: | Existing methods for unlearning large language models fine-tune by maximizing loss, but they are unstable . this creates instability, especially on larger datasets, leading to over-unlearning . |
| Approach: | They propose a novel unlearning method that leverages self-distillation to adjust logits . this method ensures smooth convergence and avoids catastrophic forgetting . |
| Outcome: | The proposed method achieves smooth convergence and avoids catastrophic forgetting even on large datasets and sequential unlearning requests. |
Confidence Estimation for LLMs in Multi-turn Interactions (2026.findings-acl)
Copied to clipboard
Caiqi Zhang, Ruihan Yang, Xiaochen Zhu, Chengzu Li, Tiancheng Hu, Yijiang River Dong, Deqing Yang, Nigel Collier
| Challenge: | Despite recent progress, most prior work studies confidence in single-turn question answering. |
| Approach: | They propose a logit-based probe that measures confidence in multi-turn dialogues . they propose 'infoECE' and a "hinter-guesser" paradigm for generating controlled evaluations based on data . |
| Outcome: | The proposed framework is grounded in calibration and monotonicity of confidence as more information becomes available. |
Privacy-R1: Privacy-Aware Multi-LLM Agent Collaboration via Reinforcement Learning (2026.acl-long)
Copied to clipboard
| Challenge: | Prior approaches to rewriting large language models shatters linguistic coherence and removes privacy-sensitive information. |
| Approach: | They propose a framework that trains an agent to dynamically route text chunks . it implicitly distinguishes between replaceable Personally Identifiable Information (PII) and task-critical PII . |
| Outcome: | The proposed framework achieves state-of-the-art on the privacy-utility frontier . it trains an agent to dynamically route text chunks, learning a policy that balances privacy leakage and task performance. |
Value of Information: A Framework for Human–Agent Communication (2026.acl-long)
Copied to clipboard
| Challenge: | Existing approaches to large language model (LLM) agents fail to account for stakes of different decisions. |
| Approach: | They propose a framework that balances task risk, query ambiguity, user effort . they use a value-of-information framework to dynamically weigh the expected utility gain . |
| Outcome: | The proposed model matches or exceeds the best manually-tuned baselines in four domains . it explicitly balances task risk, query ambiguity, and user effort . |