First author of 'SEED: Accelerating Reasoning Tree Construction via Scheduled Speculative Decoding', accepted by COLING 2025 and NeurIPS 2024 Compression Workshop
Co-first author of 'SCOPE: Optimizing Key-Value Cache Compression in Long-context Generation', accepted by ACL 2025 main conference
Co-author of 'WebWalker: Benchmarking LLMs in Web Traversal', accepted by ACL 2025 main conference
First author of preprint 'Benchmarking Temporal Reasoning and Alignment Across Chinese Dynasties'
Co-author of 'Opinions Are Not Always Positive: Debiasing Opinion Summarization with Model-Specific and Model-Agnostic Methods', published at LREC-COLING 2024
Background
M.S. student in Artificial Intelligence at Southeast University
Former member of the PAttern Learning and Mining (PALM) Lab
Research focuses on Efficient LLMs Inference and Text Generation
Currently working on KV Cache Compression and Long-context LLMs Inference