ICLR

UBC Computer Science publishes 21 papers at top international machine learning conference

A total of 9 papers were accepted at the International Conference on Learning Representation (ICLR) main conference and 12 papers at the associated workshops 

Researchers from UBC Computer Science will be in Rio de Janeiro, Brazil from April 23 - 27, 2026 for the 14th International Conference on Learning Representations (ICLR). Each year, thousands of computer science researchers and industry leaders from around the world gather at ICLR to present the latest research and discuss breakthroughs in the field. 

Nearly 20,000 papers were submitted to the conference and 5,355 papers in total were accepted. UBC Computer Science had 9 papers accepted at the main conference and 12 papers accepted to the various workshops.  

A paper from Assistant Professor Mi Jung Park’s group on safe image generation was accepted as an oral presentation, representing the top 1% of conference paper submissions.  

In addition to papers at the conference and workshops, Assistant Professor Evan Shelhamer is Co-Chair of mentoring and part of the organizing team for the Machine Learning for Remote Sensing Workshop and the Test-Time Updates Workshop.  

The 9 papers at the main conference are: 

  1. DRBench: A Realistic Benchmark for Enterprise Deep Research 
    Amirhossein Abaskohi, Tianyi Chen, Miguel Muñoz-Mármol, Curtis Fox, Amrutha Varshini Ramesh, Etienne Marcotte, Xing Han , Nicolas Chapados, Spandana Gella, Peter West, Giuseppe Carenini, Christopher Pal, Alexandre Drouin, Issam H. Laradji 
     

  1. Physically Valid Biomolecular Interaction Modeling with Gauss-Seidel Projection 
    Siyuan Chen, Minghao Guo, Caoliwen Wang, Anka He Chen, Yikun Zhang, Jingjing Chai, Yin Yang, Wojciech Matusik, Peter Yichen Chen 
     

  1. SysMoBench: Evaluating AI on Formally Specifying Complex Real-World Systems 
    Qian Cheng, Ruize Tang, Emilie Ma, Finn Hackett, Peiyang He, Yiming Su, Ivan Beschastnikh, Yu Huang, Xiaoxing Ma, Tianyin Xu 
     

  1. Thicker and Quicker: The Jumbo Token for Fast Plain Vision Transformers 
    Anthony Fuller, Yousef Yassin, Daniel Kyrollos, Evan Shelhamer, James Green 
     

 

UBC Computer Science also has 12 accepted papers in various ICLR workshops: 

  1. Bridging Generative and Predictive Paradigms via Hidden-Self-Distillation 
    Scott C. Lowe, Anthony Fuller, Sageev Oore, Evan Shelhamer, Graham W. Taylor 
    Multimodal Intelligence: Next Token Prediction & Beyond Workshop 
     

  1. Changing Modalities by Cross-Band Transfer, Addition, and Peeking 
    Tim G. Zhou, Anthony Fuller, Geoff Pleiss, Evan Shelhamer 
    Machine Learning for Remote Sensing Workshop (ML4RS)  
     

  1. Discrete Meanflow Training Curriculum 
    Chia-Hong HSU, Frank Wood 
    Deep Generative Model in Machine Learning: Theory, Principle and Efficacy Workshop (DeLTa) 
     

  1. Hidden-Layer Self-Distillation Yields Drift-Resilient Visual Representations 
    Scott C. Lowe, Anthony Fuller, Sageev Oore, Graham W. Taylor, Evan Shelhamer 
    Catch, Adapt, and Operate: Monitoring ML Models Under Drift Workshop (CAO) 
     

  1. Learning to Continually Learn via Meta-learning Agentic Memory Designs 
    Yiming Xiong, Shengran Hu, Jeff Clune 
    AI with Recursive Self-Improvement Workshop (RSI) & 
    Memory for LLM-Based Agentic Systems Workshop 
     

  1. LookSharp: Attention Entropy Minimization for Test-Time Adaptation 
    Yash MaliEvan Shelhamer 
    Catch, Adapt, and Operate: Monitoring ML Models Under Drift Workshop (CAO) & 
    Test-Time Updates Workshop (TTU) 
     

  1. ProtoTTA: Prototype-Guided Test-Time Adaptation 
    Mohammad Mahdi Abootorabi, Parvin Mousavi, Purang Abolmaesumi, Evan Shelhamer 
    Test-Time Updates Workshop (TTU) 
     

  1. RDUMB++: Drift-Aware Continual Test-time Adaptation 
    Himanshu Mishra 
    Catch, Adapt, and Operate: Monitoring ML Models Under Drift Workshop (CAO) & 
    Test-Time Updates Workshop (TTU) 
     

  1. Self-Soupervision: Cooking and Seasoning Model Soups without Labels for Adaptation 
    Anthony Fuller, James R Green, Evan Shelhamer 
    Test-Time Updates Workshop (TTU) 
     

  1. Value Drifts: Tracing Value Alignment During LLM Post-Training 
    Mehar Bhatia, Shravan Nayak, Gaurav Kamath, Marius Mosbach, Karolina Stanczak, Vered Shwartz, Siva Reddy 
    Catch, Adapt, and Operate: Monitoring ML Models Under Drift Workshop (CAO)