- [Aug, 2026] Two first-authored papers are accepted by EMNLP 2026 (Main). GPAgentBench-2K for benchmarking LLM agents on sequential decision-making in a complex action space. Revisiting Greedy Decoding for VQA for rethinking multimodal LLM decoding from a calibration perspective.
- [Jun, 2026] My co-first-authored paper Controllable Egocentric Video Generation via Occlusion-Aware Sparse 3D Hand Joints is accepted by ECCV 2026. It generates egocentric videos controlled by sparse 3D hand joints and transfers to robotic hands.
- [Apr, 2026] My first-authored paper MEDSYN is accepted by ACL 2026 Findings. It benchmarks how multimodal LLMs synthesize heterogeneous evidence in complex cases.
- [Jan, 2026] I joined Apple in Shanghai as a Machine Learning & AI Engineer Intern, working on LLM agents for failure analysis.
- [Jan, 2026] My co-authored paper YoNoSplat is accepted by ICLR 2026. It is a single feedforward model for 3D Gaussian Splatting from posed or unposed images.
- [May, 2025] My first-authored paper Revisiting Automatic Data Curation for Vision Foundation Models in Digital Pathology is early accepted by MICCAI 2025 (top 9%). It curates web-scale data for foundation model pretraining.
- [Oct, 2024] My first-authored paper ICMSeg is accepted by WACV 2025 for single-source cross-modality medical image segmentation via invariant causal mechanisms.
- [Jun, 2023] I started my PhD in Computer Science at ETH Zurich as an ETH AI Center Doctoral Fellow.
- [Jul, 2022] My co-first-authored paper ZoomMIL is accepted by ECCV 2022. It learns differentiable zooming for multiple instance learning on whole-slide images.
- [Sep, 2021] I joined IBM Research Zurich as a research intern in the Accelerated Discovery & AI group.