Biography

Hi, I’m Keyu, a second-year master student in Machine Learning at University of Tuebingen. Currently, I’m a research assistant in Max Planck Institute for Intelligent Systems, working closely with Dr. Shiwei Liu and Dr. Jonas Geiping. I earned my B.Eng. in Artificial Intelligence from Southeast University where I was fortunate to be advised by Prof. Guilin Qi, and had a wonderful time at SEU-131AIClub, contributing to its founding and flourishing.

I am passionate about making AI scalable. My current research centers on scalable and efficient foundation models, with a focus on understanding and enhancing their architectural design, reasoning capabilities, training and data efficiency, and system reliability.

You can find my CV here. Please feel free to reach out to me via email at keyu.wang@student.uni-tuebingen.de.

Education

2024.10 - Present, M.Sc. in Machine Learning, University of Tuebingen, Tuebingen, Germany
2020.09 - 2024.06, B.Eng. in Artificial Intelligence, Southeast University, Nanjing, China

Research Experience

2025.10 - Present, SEAL Lab, Max Planck Institute for Intelligence Systems & ELLIS Tuebingen. Supervisor: Dr. Jonas Geiping
2025.03 - Present, WEI Lab, Max Planck Institute for Intelligence Systems & ELLIS Tuebingen. Supervisor: Dr. Shiwei Liu
2022.02 - 2024.06, Knowledge Science and Engineering Lab, Southeast University. Supervisor: Prof. Dr. Guilin Qi

Industrial Internship

2023.07 - 2024.06, intern, BSH Home Appliances Holding (China) Co., Ltd, Nanjing, China

Publications & Preprints

Refer to my Google Scholar for a complete list. (* Equal contribution)

I. Model Scalability

When Fewer Layers Break More Chains: Layer Pruning Harms Test-Time Scaling in LLMs

Keyu Wang*, Tian Lyu*, Guinan Su, Jonas Geiping, Lu Yin, Marco Canini, Shiwei Liu
ArXiv 2025 [PDF]

Demystifying the Roles of LLM Layers in Retrieval, Knowledge, and Reasoning

Xinyuan Song, Keyu Wang, Pengxiang Li, Lu Yin and Shiwei Liu
ICASSP 2026 [PDF]

II. Training Scalability

One LR Doesn’t Fit All: Heavy-Tail Guided Layerwise Learning Rates for LLMs

Di He, Songjun Tu, Keyu Wang, Lu Yin and Shiwei Liu
ICLR 2026 DeLTa Workshop Poster

III. Reasoning/Inference Scalability

Can Large Language Models Understand DL-Lite Ontologies? An Empirical Study

Keyu Wang, Guilin Qi, Jiaqi Li and Songlin Zhai
EMNLP 2024 Findings [PDF]

An Embedding-based Approach to Inconsistency-tolerant Reasoning with Inconsistent Ontologies

Keyu Wang, Site Li, Jiaye Li, Guilin Qi, and Qiu Ji
IJCKG 2023 [PDF]

Academic Services

Reviewer: CPAL 2026

Misc

My Chinese name is 汪可予. I come from Tongling, Anhui — a small city by the Yangtze River.
I like all kinds of sports, such as table tennis, basketball, football, badminton, marathons, jogging, hiking and climbing.
I a big fan of animations. I enjoy various chess, board and card games.