Chang Liu 刘昶

I am a Ph.D. student in the Machine Learning Department at Carnegie Mellon University and a member of the Auton Lab, advised by Prof. Artur Dubrawski. I did my undergraduate studies in the Yao Class at Tsinghua University.

My research spans machine learning for healthcare and LLM post-training. My work aims to democratize expertise: on the healthcare side, I develop affordable outbreak detection systems for hospitals using multimodal ML on routinely collected data (MALDI-TOF spectra, AMR profiles and EHR); on the LLM side, I design learning scaffolds for LLM post-training, inspired by cognitive learning strategies of humans.

Portrait of Chang Liu

Education

  • 2023 — 2028 (est.)Pittsburgh, US

    Ph.D., Machine Learning

    Carnegie Mellon University, School of Computer Science

  • 2019 — 2023Beijing, China

    B.Eng., Computer Science

    Tsinghua University — Yao Class, Institute for Interdisciplinary Information Sciences · GPA 3.91/4.00

Industry

  • 05/2026 — 08/2026San Jose, US

    Research Intern, Video Recommendation

    TikTok

    Modeling user preference over professionally generated content (PGC) to improve recommendation performance.

Publications

  • EMNLP FindingsAug 2026

    Notes to Self: Can LLMs Benefit from Experiential Abstractions?

    Chang Liu, Xinyu Li, Artur Dubrawski*

    [link]

  • AMIA Annual SymposiumMay 2026

    Towards Practical Multimodal Hospital Outbreak Detection

    Chang Liu, Jieshi Chen, Alexander J. Sundermann, Kathleen Shutt, Marissa P. Griffith, Lora Lee Pless, Lee H. Harrison, Artur Dubrawski*

    [link]

  • AMIA Amplify Informatics ConferenceFeb 2026

    Exploring the Utility of MALDI-TOF Mass Spectrometry and Antimicrobial Resistance in Hospital Outbreak Detection

    Chang Liu, Jieshi Chen, Alexander J. Sundermann, Kathleen Shutt, Marissa P. Griffith, Lora Lee Pless, Lee H. Harrison, Artur Dubrawski*

    [link]

  • NeurIPS Workshop · LLM EvalSep 2025

    Depth as a Scaling Vector: Simple Pruning and Evaluation of Emergent Abilities in Pruned LLMs

    Chang Liu†, Arjun Choudhry†, Yifu Cai, Nina Żukowska, Mononito Goswami, Artur Dubrawski

    [link]

  • NeurIPS Workshop · Efficient ReasoningSep 2025

    LayerMerge: Modality-Agnostic Depth Pruning for Efficient Foundation Model Deployment

    Arjun Choudhry†, Chang Liu†, Nina Żukowska, Yifu Cai, Mononito Goswami, Artur Dubrawski

    [link]

  • NeurIPS Workshop · UniRepsSep 2025

    From Aggregation to Guidance: Strategies for Personalized Federated Fine-Tuning of Foundation Models

    Mikołaj Piórczyński, Wojciech Łapacz, Xinyu Li, Chang Liu, Abby Turner, Artur Dubrawski

    [link]

  • CHILApr 2025

    Bridging the Utility Gap Between MALDI-TOF and WGS for Affordable Outbreak Cluster Detection

    Chang Liu, Jieshi Chen, Lee H. Harrison, Artur Dubrawski*

    [link]

  • arXiv preprintOct 2024

    Multimodal Structure Preservation Learning

    Chang Liu, Jieshi Chen, Lee H. Harrison, Artur Dubrawski*

    [link]

  • PLOS Computational BiologyApr 2024

    A Probabilistic Knowledge Graph for Target Identification

    Chang Liu†, Kaimin Xiao†, Cuinan Yu†, Yipin Lei†, … Dan Zhao*, Fengfeng Zhou*, Haidong Tang*, Jianyang Zeng*

    [link]

  • Pacific Symposium on BiocomputingJan 2023

    Improving Target-disease Association Prediction through a Graph Neural Network with Credibility Information

    Chang Liu†, Cuinan Yu†, Yipin Lei†, … Dan Zhao*, Fengfeng Zhou*, Jianyang Zeng*

    [link]

† equal contribution  ·  * corresponding author

Teaching

  • Fall 2025

    10-417/617 Intermediate Deep Learning

    TA for Prof. Ruslan Salakhutdinov

  • Spring 2025

    10-707 Advanced Deep Learning

    TA for Prof. Ruslan Salakhutdinov

Personal

I am a skilled and avid pianist with a wholehearted love for classical music (e.g., Chopin's). See my recital here. I am currently working on Chopin's Preludes (Op. 28) and Sonata in B minor (Op. 58).