Selected Publications

† Corresponding author * Equal Contribution

2026

• The Geometry of Agent Skills: Non-Commutative Composition in Representation Space
  Junda Wu, Yifan Wang, Zihan Huang, Xunyi Jiang, Sheldon Yu, Rohan Surana, Lina Yao, Julian McAuley, Tong Yu.
  NeurIPS 2026 [Paper]

• Reasoning Pathologies in Large Language Models: A Diagnostic Perspective
  Rohan Surana, Junda Wu, Sheldon Yu, Gagan Mundada, Xunyi Jiang, Zihan Huang, Ryan A. Rossi, Julian McAuley, Tong Yu.
  NeurIPS 2026 [Paper]

• How Hard Does It Think? Analyzing Step-Aware Reasoning Energy in LLM Chain-of-Thought Trajectories
  Hui Wei, Junda Wu, Sheldon Yu, Sizhe Zhou, Yizhu Jiao, Ming Zhong, Bowen Jin, Tong Yu, Shijia Pan, Jiawei Han, Julian McAuley.
  arXiv 2026 [Paper]

• RRPO: Reference-Relative Policy Optimization with Stratified Conditional Rollouts
  Yuxin Xiong, Xunyi Jiang, Rohan Surana, Xintong Li, Sheldon Yu, Nikki Lijing Kuang, Ryan A. Rossi, Jingbo Shang, Tong Yu, Julian McAuley, Junda Wu.
  arXiv 2026 [Paper]

• Can We Break LLMs Out of Self-Loops? Fine-Grained Reasoning Control with Activation Steering
  Sheldon Yu, Tong Yu, Xunyi Jiang, Rohan Surana, Gagan Mundada, Sungchul Kim, Lina Yao, Julian McAuley, Junda Wu.
  arXiv 2026 [Paper]

• Toward Latent Language Model Skills Steering and Optimization: An Empirical Study
  Xunyi Jiang, Junda Wu, Yuxin Xiong, Sheldon Yu, Tong Yu, David Arbour, Ritwik Sinha, Julian McAuley, Hongyi Wen.
  EMNLP 2026

• BELIEFMEM: Belief-Revising Memory for Conversational LLM Agents
  Felix Meng, Junda Wu, Julian McAuley.
  EMNLP 2026 Findings

• An Empirical Study on Zero-Data Bootstrapping for Conversational Recommender Systems
  Rohan Surana, Junda Wu, Zhouhang Xie, Yu Xia, Nathan Kallus, Julian McAuley.
  EMNLP 2026 Findings

• STARS: State-Aware Reasoning Stopping via Hidden-State Spectral Analysis
  Aaron Cheng, Junda Wu, Sheldon Yu, Rui Wang, Ryan A. Rossi, Sizhe Zhou, Jiawei Han, Julian McAuley, Lina Yao, Tong Yu.
  EMNLP 2026 Findings

• SymSkill: Skill Graph for Symbolic Music Understanding
  Junda Wu, Boyang Wang, Xunyi Jiang, Aditya Yadavalli, Ziyu Wang, Qianqi Yan, Hongyi Wen, Julian McAuley.
  ISMIR 2026

• MixWeaver: Stem-Aware Automatic DJ Transition Generation via Structured Discrete Optimization
  Xuran Zhou, Xunyi Jiang, Churan Zhi, Junda Wu, Julian McAuley.
  ISMIR 2026

• Multi-View Structural Interpretability of Agentic Reasoning Traces
  Ritika Hooda, Sravanthi Machcha, Jackson MacDonald, Lavanika Srinivasaraghavan, Atharva Nijasure, Rui Wang, Lei Shi, Junda Wu, Tong Yu.
  EMNLP 2026 Industry

• Skill-CMIB: Multimodal Agent Skill for Consistent Action via Conditional Multimodal Information Bottleneck
  Zihan Huang*, Junda Wu*, Tong Yu, Qianqi Yan, Rohan Surana, Uttaran Bhattacharya, Lina Yao, Xin Eric Wang, Julian McAuley.
  arXiv 2026 [Paper]

• Skill-R1: Agent Skill Evolution via Reinforcement Learning
  Yash Vishe, Rohan Surana, Xunyi Jiang, Zihan Huang, Xintong Li, Nikki Lijing Kuang, Tong Yu, Ryan A Rossi, Jingbo Shang, Julian McAuley, Junda Wu*.
  arXiv 2026 [Paper]

• OLIVIA: Online Learning via Inference-time Action Adaptation
  Sheldon Yu*, Junda Wu*, Xintong Li, Nikki Kuang, Sizhe Zhou, Tong Yu, Jiawei Han, Jingbo Shang, Julian McAuley.
  EMNLP 2026 [Paper]

• Ctrls: Chain-of-thought reasoning via latent state-transition
  Junda Wu, Yuxin Xiong, Xintong Li, Sheldon Yu, Zhengmian Hu, Tong Yu, Rui Wang, Xiang Chen, Jingbo Shang, Julian McAuley.
  AISTATS 2026 [Paper]

• Musicrs: Benchmarking audio-centric conversational recommendation
  Rohan Surana, Amit Namburi, Gagan Mundada, Abhay Lal, Zachary Novack, Julian McAuley, Junda Wu†.
  ICASSP 2026 [Paper] [Code]

• WS-GRPO: Weakly-Supervised Group-Relative Policy Optimization for Rollout-Efficient Reasoning
  Gagan Mundada, Zihan Huang, Rohan Surana, Sheldon Yu, Jennifer Yuntong Zhang, Xintong Li, Tong Yu, Lina Yao, Jingbo Shang, Julian McAuley, Junda Wu†.
  ICML 2026 [Paper]

• Generate, Filter, Control, Replay: A Comprehensive Survey of Rollout Strategies for LLM Reinforcement Learning
  Rohan Surana, Gagan Mundada, Xunyi Jiang, Chuhan Wang, Zhenwei Tang, Difan Jiao, Zihan Huang, Yuxin Xiong, Junda Wu, Sheldon Yu, Xintong Li, Raghav Jain, Nikki Kuang, Sizhe Zhou, Bowen Jin, Zhendong Chu, Tong Yu, Ryan A. Rossi, Kuan-Hao Huang, Jingbo Shang, Jiawei Han, Julian McAuley.
  arXiv 2026 [Paper]

• AMPS: Adaptive Modality Preference Steering via Functional Entropy
  Zihan Huang, Xintong Li, Rohan Surana, Tong Yu, Rui Wang, Julian McAuley, Jingbo Shang, Junda Wu†.
  NeurIPS 2026 [Paper]

• MASS-DPO: Multi-negative Active Sample Selection for Direct Policy Optimization
  Rohan Surana*, Junda Wu*, Xintong Li, Sheldon Yu, Yiran Jenny Shen, Chuhan Wang, Tong Yu, Prithviraj Ammanabrolu, Jingbo Shang, Julian McAuley.
  arXiv 2026 [Paper]

• Importance Sampling for Multi-Negative Multimodal Direct Preference Optimization
  Xintong Li, Chuhan Wang, Junda Wu, Rohan Surana, Tong Yu, Julian McAuley, Jingbo Shang.
  ICLR 2026 [Paper]

• CSyMR: Benchmarking Compositional Music Information Retrieval in Symbolic Music Reasoning
  Boyang Wang, Yash Vishe, Xin Xu, Zachary Novack, Xunyi Jiang, Julian McAuley, Junda Wu†.
  SIGIR 2026 [Paper]

• Dice: Dynamic in-context example selection in llm agents via efficient knowledge transfer
  Ruoyu Wang, Junda Wu, Yu Xia, Tong Yu, Ryan A. Rossi, Julian McAuley, Lina Yao.
  SIGKDD 2026 [Paper]

• CachePrune: Teaching LLMs What Not to Follow via KV-Cache Editing
  Rui Wang, Junda Wu, Yu Xia, Tong Yu, Ruiyi Zhang, Ryan A. Rossi, Subrata Mitra, Lina Yao, Julian McAuley.
  ACL 2026 [Paper]

2025

• Pdb-eval: An evaluation of large multimodal models for description and explanation of personalized driving behavior
  Junda Wu, Jessica Echterhoff, Kyungtae Han, Amr Abdelraouf, Rohit Gupta, Julian McAuley.
  IV 2025 [Paper]

• WildScore: Benchmarking MLLMs in-the-Wild Symbolic Music Reasoning
  Gagan Mundada, Yash Vishe, Amit Namburi, Xin Xu, Zachary Novack, Julian McAuley, Junda Wu†.
  EMNLP 2025 [Paper] [Code]

• CoMMIT: Coordinated Multimodal Instruction Tuning
  Xintong Li*, Junda Wu*, Tong Yu, Rui Wang, Yu Wang, Xiang Chen, Jiuxiang Gu, Lina Yao, Julian McAuley, Jingbo Shang.
  EMNLP 2025 [Paper]

• Image Difference Captioning via Adversarial Preference Optimization
  Zihan Huang*, Junda Wu*, Rohan Surana, Tong Yu, David Arbour, Ritwik Sinha, Julian McAuley.
  EMNLP 2025 [Paper]

• Mitigating Visual Knowledge Forgetting in MLLM Instruction-tuning via Modality-decoupled Gradient Descent
  Junda Wu, Yuxin Xiong, Xintong Li, Yu Xia, Ruoyu Wang, Yu Wang, Tong Yu, Sungchul Kim, Ryan A. Rossi, Lina Yao, Jingbo Shang, Julian McAuley.
  EMNLP 2025 Findings [Paper]

• IRPO: In-context ranking preference optimization
  Junda Wu*, Rohan Surana*, Zhouhang Xie, Yiran Shen, Yu Xia, Tong Yu, Ryan A. Rossi, Prithviraj Ammanabrolu, Julian McAuley.
  COLM 2025 [Paper] [Code]

• Traceable and Explainable Multimodal Large Language Models: An Information-Theoretic View
  Zihan Huang*, Junda Wu*, Rohan Surana, Raghav Jain, Tong Yu, Raghavendra Addanki, David Arbour, Sungchul Kim, Julian McAuley.
  COLM 2025 [Paper] [Code]

• Collap: Contrastive long-form language-audio pretraining with musical temporal structure augmentation
  Junda Wu, Warren Li, Zachary Novack, Amit Namburi, Carol Chen, Julian McAuley.
  ICASSP 2025 [Paper]

• FUTGA-MIR: Enhancing Fine-grained and Temporally-aware Music Understanding with Music Information Retrieval
  Junda Wu, Zachary Novack, Amit Namburi, Hao-Wen Dong, Carol Chen, Jiaheng Dai, Julian McAuley.
  ICASSP 2025

  [Code]

• Doc-React: Multi-page Heterogeneous Document Question-answering
  Junda Wu, Yu Xia, Tong Yu, Xiang Chen, Sai Sree Harsha, Akash V. Maharaj, Ruiyi Zhang, Victor Bursztyn, Sungchul Kim, Ryan A. Rossi, Julian McAuley, Yunyao Li, Ritwik Sinha.
  ACL 2025 [Paper]

• OCEAN: Offline chain-of-thought evaluation and alignment in large language models
  Junda Wu, Xintong Li, Ruoyu Wang, Yu Xia, Yuxin Xiong, Jianing Wang, Tong Yu, Xiang Chen, Branislav Kveton, Lina Yao, Jingbo Shang, Julian McAuley.
  ICLR 2025 [Paper]

• A Tutorial on Agentic LLM for Recommender Systems
  Chengkai Huang, Junda Wu, Tong Yu, Julian McAuley, Lina Yao.
  RecSys 2025 [Paper]

• A survey on personalized and pluralistic preference alignment in large language models
  Zhouhang Xie, Junda Wu, Yiran Shen, Yu Xia, Xintong Li, Aaron Chang, Ryan A. Rossi, Sachin Kumar, Raghav Jain, Tong Yu, Bodhisattwa Prasad Majumder, Jingbo Shang, Prithviraj Ammanabrolu, Julian McAuley.
  COLM 2025 [Paper]

• Knowledge-Aware Query Expansion with Large Language Models for Textual and Relational Retrieval
  Yu Xia, Junda Wu, Sungchul Kim, Tong Yu, Ryan A. Rossi, Haoliang Wang, Julian McAuley.
  NAACL 2025 [Paper]

• Towards agentic recommender systems in the era of multimodal large language models
  Chengkai Huang, Junda Wu, Yu Xia, Zixu Yu, Ruhan Wang, Tong Yu, Ruiyi Zhang, Ryan A. Rossi, Branislav Kveton, Dongruo Zhou, Julian McAuley, Lina Yao.
  ACM TIST 2025 [Paper]

• Pluralistic Off-policy Evaluation and Alignment
  Chengkai Huang, Junda Wu, Zhouhang Xie, Yu Xia, Rui Wang, Tong Yu, Subrata Mitra, Julian McAuley, Lina Yao.
  arXiv 2025 [Paper]

• From reviews to dialogues: Active synthesis for zero-shot llm-based conversational recommender system
  Rohan Surana, Junda Wu, Zhouhang Xie, Yu Xia, Harald Steck, Dawen Liang, Nathan Kallus, Julian McAuley.
  arXiv 2025 [Paper]

• Active learning for direct preference optimization
  (Alphabetical Order) Branislav Kveton, Xintong Li, Julian McAuley, Ryan A. Rossi, Jingbo Shang, Junda Wu, Tong Yu.
  arXiv 2025 [Paper]

2024

• DeCoT: Debiasing Chain-of-Thought for Knowledge-Intensive Tasks in Large Language Models via Causal Intervention
  Junda Wu, Tong Yu, Xiang Chen, Haoliang Wang, Ryan A. Rossi, Sungchul Kim, Anup B. Rao, Julian McAuley.
  ACL 2024 [Paper]

• CoRAL: Collaborative Retrieval-Augmented Large Language Models Improve Long-tail Recommendation
  Junda Wu, Cheng-Chun Chang, Tong Yu, Zhankui He, Jianing Wang, Yupeng Hou, Julian McAuley.
  SIGKDD 2024 [Paper]

• Neighborhood-based collaborative filtering for conversational recommendation
  Zhouhang Xie, Junda Wu, Hyunsik Jeon, Zhankui He, Harald Steck, Rahul Jha, Dawen Liang, Nathan Kallus, Julian McAuley.
  RecSys 2024 [Paper] [Code]

• InstructGraph: Boosting Large Language Models via Graph-centric Instruction Tuning and Preference Alignment
  Jianing Wang, Junda Wu, Yupeng Hou, Yao Liu, Ming Gao, Julian McAuley.
  ACL 2024 Findings [Paper] [Code]

• FUTGA: Towards Fine-grained Music Understanding through Temporally-enhanced Generative Augmentation
  Junda Wu, Zachary Novack, Amit Namburi, Jiaheng Dai, Hao-Wen Dong, Zhouhang Xie, Carol Chen, Julian McAuley.
  NLP4MusA 2024 [Paper] [Code]

• Personalized multimodal large language models: A survey
  Junda Wu, Hanjia Lyu, Yu Xia, Zhehao Zhang, Joe Barrow, Ishita Kumar, Mehrnoosh Mirtaheri, Hongjie Chen, Tong Yu, Ryan A. Rossi, et al..
  arXiv 2024 [Paper]

• Visual Prompting in Multimodal Large Language Models: A Survey
  Junda Wu, Zhehao Zhang, Yu Xia, Xintong Li, Zhaoyang Xia, Aaron Chang, Tong Yu, Sungchul Kim, Ryan A. Rossi, Ruiyi Zhang, Subrata Mitra, Dimitris N. Metaxas, Lina Yao, Jingbo Shang, Julian McAuley.
  arXiv 2024 [Paper]

2023

• InfoPrompt: Information-Theoretic Soft Prompt Tuning for Natural Language Understanding
  Junda Wu, Rui Wang, Tong Yu, Zhao Song, Ruiyi Zhang, Handong Zhao, Chaochao Lu, Shuai Li, Ricardo Henao.
  NeurIPS 2023 [Paper] [Code]

• User-Regulation Deconfounded Conversational Recommender System with Bandit Feedback
  Yu Xia, Junda Wu, Tong Yu, Sungchul Kim, Ryan A. Rossi, Shuai Li.
  KDD 2023 [Paper]

• Content-aware progressive image compression and syncing
  Junda Wu, Haoliang Wang, Tong Yu, Gang Wu, Stefano Petrangeli, Handong Zhao, Sungchul Kim, Viswanathan Swaminathan.
  ISM 2023

• Few-Shot Composition Learning for Image Retrieval with Prompt Tuning
  Junda Wu, Rui Wang, Handong Zhao, Ruiyi Zhang, Chaochao Lu, Shuai Li, Ricardo Henao.
  AAAI 2023 [Paper] [Code]

2022

• Context-aware Information-theoretic Causal De-biasing for Interactive Sequence Labeling
  Junda Wu, Rui Wang, Tong Yu, Ruiyi Zhang, Handong Zhao, Shuai Li, Ricardo Henao, Ani Nenkova.
  EMNLP 2022 Findings [Paper]

• Dynamics-aware adaptation for reinforcement learning based cross-domain interactive recommendation
  Junda Wu, Zhihui Xie, Tong Yu, Handong Zhao, Ruiyi Zhang, Shuai Li.
  SIGIR 2022 [Paper]

2021

• Deconfounded and explainable interactive vision-language retrieval of complex scenes
  Junda Wu, Tong Yu, Shuai Li.
  ACM MM 2021 [Paper]

• Clustering of conversational bandits for user preference learning and elicitation
  Junda Wu, Canzhe Zhao, Tong Yu, Jingyang Li, Shuai Li.
  CIKM 2021 [Paper]

• Sim-to-real interactive recommendation via off-dynamics reinforcement learning
  Junda Wu, Zhihui Xie, Tong Yu, Qizhi Li, Shuai Li.
  NeurIPS 2021 Workshop [Paper]

US Patents

• Progressive image compression and syncing
  Junda Wu, Haoliang Wang, Tong Yu, Stefano Petrangeli, Gang Wu, Viswanathan Swaminathan, Sungchul Kim, Ryan A. Rossi.
  US Patent 12,579,699, 2026

• Dialogue state aware dialogue summarization
  Haoliang Wang, Kaige Xie, Tong Yu, Junda Wu, Handong Zhao, Ruiyi Zhang, Kanak Mahadik, Ani Nenkova.
  US Patent 12,566,925, 2026

• Transferable clustering of contextual bandits for cloud service resource allocation
  Kanak Mahadik, Tong Yu, Junda Wu.
  US Patent 12,294,529, 2025

• Chain-of-thought machine-learning model debiasing
  Haoliang Wang, Xiang Chen, Tong Yu, Sungchul Kim, Ryan A. Rossi, Junda Wu, Anup B. Rao.
  US Patent App. 18/673,547, 2025

• Dialogue skeleton assisted prompt transfer for dialogue summarization
  Tong Yu, Kaige Xie, Junda Wu, Handong Zhao, Ruiyi Zhang, Kanak Mahadik, Ani Nenkova.
  US Patent App. 18/355,901, 2025

• Contextual query generation
  Haoliang Wang, Tong Yu, Sungchul Kim, Ruiyi Zhang, Peng Xu, Junda Wu, Handong Zhao, Ani Nenkova.
  US Patent App. 18/339,694, 2024