UMD Researchers Present 16 Papers at ICML 2025

Researchers from the University of Maryland’s Department of Computer Science are presenting 16 papers at the 42nd International Conference on Machine Learning (ICML 2025), which will take place from July 13 to 19 in Vancouver, Canada.

Many of the participating UMD faculty—including Furong Huang, Heng Huang, Dinesh Manocha, Ming Lin, Nirupam Roy and Tianyi Zhou—collaborated with graduate students and external partners on the work being presented. Projects include both foundational machine learning methods and application-driven approaches.

The work explores a wide range of topics in machine learning, including robustness, model interpretability, human-AI collaboration and multimodal learning. 

“This strong showing at ICML highlights the breadth of machine learning research taking place across our department,” said Department Chair Matthias Zwicker, who holds the Elizabeth Iribe Chair for Innovation and the Phillip H. and Catherine C. Horvitz Professorship. “Our faculty and students are addressing core challenges in the field while exploring new directions with broad potential impact.”

UMD Affiliated papers being presented at ICML 2025:

• De-mark: Watermark Removal in Large Language Models
  Ruibo Chen, Yihan Wu, Junfeng Guo, Heng Huang
   
• Revisiting Convergence: Shuffling Complexity Beyond Lipschitz Smoothness
  Qi He, Peiran Yu, Ziyi Chen, Heng Huang
   
• Time-Aware World Model for Adaptive Prediction and Control
  Anh Nhu, Sanghyun Son, Ming Lin
   
• Adaptive Sensitivity Analysis for Robust Augmentation against Natural Corruptions in Image Segmentation
  Laura Zheng, Wenjie Wei, Tony Wu, Jacob Clements, Shreelekha Revankar, Andre Harrison, Yu Shen, Ming Lin
   
• Audio Flamingo 2: An Audio-Language Model with Long-Audio Understanding and Expert Reasoning Abilities
  Sreyan Ghosh, Zhifeng Kong, Sonal Kumar, S Sakshi, Jaehyeon Kim, Wei Ping, Rafael Valle, Dinesh Manocha, Bryan Catanzaro
   
• Bounded Rationality for LLMs: Satisficing Alignment at Inference-Time
  Mohamad Chehade, Soumya Suvra Ghosal, Souradip Chakraborty, Avinash Reddy, Dinesh Manocha, Hao Zhu, Amrit Singh Bedi
   
• SING: Spatial Context in Large Language Model for Next-Gen Wearables
  Ayushi Mishra, Yang Bai, Priyadarshan Narayanasamy, Nakul Garg, Nirupam Roy
   
• R2-T2: Re-Routing in Test-Time for Multimodal Mixture-of-Experts
  Zhongyang Li, Ziyue Li, Tianyi Zhou
   
• Tilted Sharpness-Aware Minimization
  Tian Li, Tianyi Zhou, Jeff Bilmes
   
• Preference Controllable Reinforcement Learning with Advanced Multi-Objective Optimization
  Yucheng Yang, Tianyi Zhou, Mykola Pechenizkiy, Meng Fang
   
• Quantifying Prediction Consistency Under Fine-tuning Multiplicity in Tabular LLMs
  Faisal Hamman, Sachindra P Dissanayake, Saumitra Mishra, Freddy Lecue, Sanghamitra Dutta
   
• Speculate, then Collaborate: Fusing Knowledge of Language Models during Decoding
  Ziyao Wang, Muneeza Azmat, Ang Li, Raya Horesh, Mikhail Yurochkin
   
• Should Decision-Makers Reveal Classifiers in Online Strategic Classification?
  Han Shao, Shuo Xie, Kunhe Yang
   
• Attention-Only Transformers via Unrolled Subspace Denoising
  Peng Wang, Yifu Lu, Yaodong Yu, Druv Pai, Qing Qu, Yi Ma
   
• Scaling Laws in Patchification: An Image Is Worth 50,176 Tokens and More
  Feng Wang, Yaodong Yu, Wei Shao, Yuyin Zhou, Alan Yuille, Cihang Xie
   
• MIB: A Mechanistic Interpretability Benchmark
  Aaron Mueller, Atticus Geiger, Sarah Wiegreffe, Dana Arad, Iván Arcuschin, Adam Belfki, Yik Siu Chan, Jaden Fiotto-Kaufman, Tal Haklay, Michael Hanna, Jing Huang, Rohan Gupta, Yaniv Nikankin, Hadas Orgad, Nikhil Prakash, Anja Reusch, Aruna Sankaranarayanan, Shun Shao, Alessandro Stolfo, Martin Tutek, Amir Zur, David Bau, Yonatan Belinkov

—Story by Samuel Malede Zewdu, CS Communications