news

Jan 26, 2026 One paper accepted to ICLR 2026.
FaSTA$^*$: Fast-Slow Toolpath Agent with Subroutine Mining for Efficient Multi-turn Image Editing
Sep 18, 2025 One paper accepted to NeurIPS 2025. See you in San Diego! 🏖️
ColorBench: Can VLMs See and Understand the Colorful World? A Comprehensive Benchmark for Color Perception, Reasoning, and Robustness
Jul 07, 2025 One paper accepted to COLM 2025. See you in Montreal! 🇨🇦
DynaSaur 🦖: Large Language Agents Beyond Predefined Actions
May 15, 2025 Our survey on GUI Agents is accepted to ACL 2025 (Findings).
Apr 10, 2025 I will join Amazon (NYC) as an Applied Scientist Intern this summer! 🍎🏙🚕🗽🐀
Jan 22, 2025 One short paper accepted to NAACL 2025.
RuleR: Improving LLM Controllability by Rule-based Data Recycling
May 16, 2024 One paper accepted to ACL 2024 (Findings).
Multi-Objective Linguistic Control of Large Language Models
Mar 24, 2024 One paper accepted to AAAI 2024.
COMBAT: Alternated Training for Effective Clean-Label Backdoor Attacks
Feb 13, 2024 I will join Adobe (San Jose) as a Research Scientist/Engineer Intern this summer!
Dec 15, 2022 One paper accepted to EMNLP 2022. We show the first empirical evidence that the off-manifold conjecture holds for natural language adversarial examples, and then propose a defense mechanism that projects adversarial examples back onto the manifold of natural sentence embeddings before classification.