Yicheng Zhang

Undergraduate Student, Computer Science and Technology
Tsinghua University, Beijing

About

I am an undergraduate student in the Department of Computer Science and Technology at Tsinghua University. I am interested in making large models faster and cheaper, at both the architecture and the systems level, and in AI systems that can help build better AI.

News

Publications

* equal contribution

  1. MLS-Bench: A Holistic and Rigorous Assessment of AI Systems on Building Better AI

    Bohan Lyu*, Yucheng Yang*, Siqiao Huang*, Jiaru Zhang*, Qixin Xu*, Yicheng Zhang*, et al.

    ICML 2026 AI for Math Workshop, 2026 Honorable Mention Award

    A benchmark that evaluates LLM agents on real ML research tasks, adopted by Kimi-K2.7-Code, Kimi-K3 and Qwen3.8-Max. I led the ML Systems & Efficient ML track.

  2. Second-Look Prompting Improves Vision-Language Models

    Yicheng Zhang, Xun Zhang, Jingyuan Ma, Tianyu Yu, Yuan Yao

    Preprint, 2026

    A training-free method that repeats the image–question sequence to enable question-aware visual processing, improving all 12 evaluated VLMs across 12 reasoning and perception benchmarks.

  3. A Survey of Efficient Attention Methods: Hardware-efficient, Sparse, Compact, and Linear Attention

    Jintao Zhang, Rundong Su, Chunyu Liu, Jia Wei, …, Yicheng Zhang, et al.

    Under review at IEEE TPAMI, 2025

    A taxonomy of efficient attention from four perspectives: hardware efficiency, sparsity, compactness and linearity. I wrote the sparse attention section.

Honors & Awards

Education

Activities