About Me
I am completing my Ph.D. in Computer Science at ETH Zurich, advised by Mrinmaya Sachan and Antoine Bosselut. Previously, I completed my M.Phil. at CUHK and my B.Eng. at HUST.
Research
My research asks how language models ground language in perceptual and contextual evidence, reason over that evidence, and remain reliable when evidence or tasks change. I combine behavioral diagnostics with mechanistic analysis to identify shortcut-driven failures, then develop training- and inference-time interventions that improve evidence use and interpretability. Multimodal settings provide a revealing testbed because they expose the interface between perception, language, and reasoning.
Selected Publications & Preprints
A complete list is available in my CV and on Google Scholar. * denotes equal contribution.
- Unveiling the Visual Counting Bottleneck in Vision-Language Models
Xingzhou Pang*, Yifan Hou*, Junling Wang, Mrinmaya Sachan.
ICML, 2026. [paper] [code]
- Diversity Matters: Revisiting Test-Time Compute in Vision-Language Models
Yijie Tong*, Yifan Hou*, Shaobo Cui, Antoine Bosselut, Mrinmaya Sachan.
ICML, 2026. [paper] [code]
- Compose and Fuse: Revisiting the Foundational Bottlenecks in Multimodal Reasoning
Yucheng Wang*, Yifan Hou*, Aydin Javadov, Mubashara Akhtar, Mrinmaya Sachan.
ICLR, 2026. [paper] [code]
- Do Vision-Language Models Really Understand Visual Language?
Yifan Hou, Buse Giledereli, Yilei Tu, Mrinmaya Sachan.
ICML, 2025. [paper] [code]
- What Do Language Models Learn in Context? The Structured Task Hypothesis
Jiaoda Li*, Yifan Hou*, Mrinmaya Sachan, Ryan Cotterell.
ACL, 2024. [paper] [code]
- Towards a Mechanistic Interpretation of Multi-Step Reasoning Capabilities of Language Models
Yifan Hou, Jiaoda Li, Yu Fei, Alessandro Stolfo, Wangchunshu Zhou, Guangtao Zeng, Antoine Bosselut, Mrinmaya Sachan.
EMNLP, 2023. [paper] [code]
- Chimera: Diagnosing Shortcut Learning in Visual-Language Understanding
Ziheng Chi*, Yifan Hou*, Chenxi Pang, Shaobo Cui, Mubashara Akhtar, Mrinmaya Sachan.
Manuscript, 2026. [paper] [code]
Student Mentoring
I have mentored research and thesis projects with students at ETH Zurich and collaborated with students from CityU of Hong Kong, HKUST, the University of Zurich, and IIIT Bangalore. These projects have contributed to publications at ICML, ICLR, and EMNLP; a complete mentoring record is available in my CV.
Experience
- Doctoral Researcher, ETH Zurich, Sep. 2020 – Sep. 2026
- Research Intern, Meta AI, Oct. 2024 – Feb. 2025
- Visiting Researcher, EPFL, Jun. – Sep. 2022 and Jun. – Sep. 2023
- Research Intern, Tencent, Jun. – Sep. 2021
- Remote Research Intern, TTIC, Jun. – Sep. 2020
- Visiting Researcher, NUS, Jun. – Sep. 2019
Selected Awards
Academic Service
Area Chair: NeurIPS; ACL Rolling Review (ARR).
Reviewer: ICML, ICLR, and COLM.
Return to Home