AI Researcher · Incoming MSc AI Student

Building AI systems that are capable, efficient, and trustworthy.

Hello, I’m Chiwun (Christian) Yang (杨智桓; Yang Chi Wun in Cantonese). I work across machine learning theory and real-world AI systems, with a focus on large language models, efficient training and inference, AI security, and learning dynamics.

In September 2026, I will begin the MSc in Artificial Intelligence at City University of Hong Kong. I received my BEng in Artificial Intelligence from Sun Yat-sen University in 2025.

Research

I am interested in mechanism-first questions at the intersection of LLMs, optimization, and systems. I especially enjoy projects that connect mathematical structure to measurable model behavior and working implementations.

01

Efficient & Long-Context LLMs

Sparse attention, KV-cache compression, parameter-efficient learning, and systems for longer, faster inference.

02

Trustworthy & Secure AI

Copyright-aware optimization, privacy and data recovery, and mathematically auditable AI behavior.

03

Learning Dynamics & Theory

Scaling laws, optimization dynamics, generalization, and the theoretical foundations of modern architectures.

Publications & Selected Preprints

* denotes equal contribution where marked. For the complete and most current record, see Google Scholar.

  1. AAAI 2024 Timothy Chu*, Zhao Song*, and Chiwun Yang*. How to Protect Copyright Data in Optimization of Large Language Models?
  2. EMNLP 2025 Yingyu Liang*, Zhenmei Shi*, Zhao Song*, and Chiwun Yang*. Towards Infinite-Long Prefix in Transformer.
  3. ICML 2025 Jing Xiong, Jianghan Shen, Chuanyang Zheng, Zhongwei Wan, Chenyang Zhao, Chiwun Yang, Fanghua Ye, Hongxia Yang, Lingpeng Kong, and Ngai Wong. ParallelComp: Parallel Long-Context Compressor for Length Extrapolation.
  4. NeurIPS 2025 Yang Cao*, Xiaoyu Li*, Zhao Song*, and Chiwun Yang*. Efficient k-Sparse Band-Limited Interpolation with Improved Approximation Ratio.
  5. CPAL 2025 Majid Daliri*, Zhao Song*, and Chiwun Yang*. Unlock the Theory behind Scaling 1-bit Neural Networks.
  6. CPAL 2025 Yekun Ke*, Yingyu Liang*, Zhenmei Shi*, Zhao Song*, and Chiwun Yang*. Curse of Attention: A Kernel-Based Perspective for Why Transformers Fail to Generalize on Time Series Forecasting and Beyond.
  7. ICLR 2025 SLLM Workshop Yichuan Deng, Zhao Song, Jing Xiong, and Chiwun Yang. How Sparse Attention Approximates Exact Attention? Your Attention is Naturally nC-Sparse.
  8. ICLR 2025 DeLTa Workshop Yang Cao*, Zhao Song*, and Chiwun Yang*. Video Latent Flow Matching: Optimal Polynomial Projections for Video Interpolation and Extrapolation.
  9. Preprint · 2023 Yichuan Deng*, Zhao Song*, Shenghao Xie*, and Chiwun Yang*. Unmasking Transformers: A Theoretical Approach to Data Recovery via Attention Weights.
  10. Preprint · 2025 Jiangxuan Long*, Zhao Song*, and Chiwun Yang*. Theoretical Foundation of Flow-Based Time Series Generation: Provable Approximation, Generalization, and Efficiency.
  11. Preprint · 2025 Chiwun Yang. Unifying Learning Dynamics and Generalization in Transformers Scaling Law.

Education & Experience

Sep 2026 – 2027
MSc in Artificial Intelligence, City University of Hong Kong
Incoming student · Department of Computer Science
2026
Founding Engineer, Humanify
Worked on agent cognition and conversational AI, including ASR and TTS
Oct 2024 – Jan 2026
Research Intern, The University of Hong Kong
Long-context compression and efficient LLM inference with Ngai Wong’s group
Oct 2022 – Mar 2025
Research Intern, Zhao Song’s Lab
Deep learning theory, optimization, and efficient transformers
Sep 2021 – Jun 2025
BEng in Artificial Intelligence, Sun Yat-sen University
School of Artificial Intelligence
Aug 2022 – Feb 2024
Research Assistant, Shenzhen Institute of Artificial Intelligence and Robotics for Society
Applied computer vision and industrial inspection

Academic Service

Reviewer for NeurIPS 2026, AAAI (2026–2027), ICLR (2025–2026), ICML (2025–2026), COLM (2025–2026), and AISTATS 2026, with additional reviewing service for an ICLR workshop and a CVPR workshop.

Let’s talk

Ideas, criticism, and collaborations are welcome.

I am always happy to discuss research on efficient and trustworthy AI, especially work that connects theory with systems.

christiannyang37 [at] gmail [dot] com

Last updated: August 2026.