leiyo@dtu.dk
Embedded Files
Department of Engineering Technology
Technical University of Denmark
I received my Ph.D. in Computer Science (specializing in Mathematical Optimization) from the Department of Information Technology at Uppsala University in 2019. During the PhD, I interned as a visiting data scientist at The Boston Consulting Group (BCG) Gamma. After the PhD, I worked as a data scientist in Bolt and Wolt (Doordash) in the domain of on-demand logistics optimization.
Research Interests
Research Interests
My research develops mathematical optimization frameworks for trustworthy and efficient AI. I treat AI as a coupled system and producing metrics, algorithms, and deployable systems for reliable evaluation and transparent governance. My work focuses on three thrusts in AI ecosystem auditing, verification first scholarly communication systems, and faithful explanations for algorithmic fairness.
Research
Research
(See a full list of publications here)
AI Ecosystem AUDITing
AI Ecosystem AUDITing
AI Ecosystem Auditing treats heterogeneous AI ecosystems, such as models, data, and platforms, as an auditable system. We develop metrics and algorithms to quantify model uniqueness and redundancy, diagnose risks and incentives in coupled data–model markets, boost models consensus, and design robust interventions to support transparent procurement, deployment, and governance.
Selected Publications
Selected Publications
L. You, "Quantifying Model Uniqueness in Heterogeneous AI Ecosystems", preprint. [arXiv] [code]
H. Ren, Y. Xiong ✉, L. You ✉, Y. Wang, H. Xiong, and Y. Zhu, "TripleWin: Fixed-Point Equilibrium Pricing for Data-Model Coupled Markets", preprint. [arXiv] [code]
L. You ✉ and H. V. Cheng ✉, "SWAP: Sparse Entropic Wasserstein Regression for Robust Network Pruning", International Conference on Learning Representations (ICLR) 2024. [arXiv] [code]
Scholarly Communication Engineering
Scholarly Communication Engineering
Scholarly Communication Engineering treats publishing, peer review, and academic credit as a system we can design. It aims to make genuine research effort the best strategy—and prevent incentive-driven collapse under AI-enabled gaming. We founded CSPaper (https://cspaper.org) as an open research infrastructure that supports end-to-end experimentation and evaluation for this agenda.
Selected Publications
Selected Publications
L. You ✉, L. Cao, and I. Gurevych, "Preventing the Collapse of Peer Review Requires Verification-First AI", preprint. [arXiv]
L. Cao ✉, L. You ✉, and CSPaper Core Team, "CSPaper Review: Fast, Rubric-Faithful Conference Feedback", International Natural Language Generation Conference (INLG) 2025. [paper] [demo] [discussion]
Explanations and Algorithmic Fairness
Explanations and Algorithmic Fairness
Our goal is to bring explainability and fairness together. We use tools like Shapley values and counterfactuals to explain why a model makes specific predictions, making sure these explanations are accurate by considering real-world data patterns. We also use these explanations as 'health checks' to see if the model is unfairly relying on sensitive traits. Based on this diagnosis, we improve the data through methods like augmentation. This approach not only makes the model more transparent and easier to audit but also ensures it treats different groups more equally.
Selected Publications
Selected Publications
Y. Bian, L. You, Y. Sasaki, H. Maeda, and A. Igarashi, "Algorithmic Fairness: Not a Purely Technical but Socio-Technical Property", preprint. [arXiv]
L. You ✉, Y. Bian, and L. Cao, "Joint Distribution–Informed Shapley Values for Sparse Counterfactual Explanations", International Conference on Learning Representations (ICLR) 2026 [arXiv] [code] [software].
L. Zhu, Y. Bian, and L. You ✉, "FairSHAP: Preprocessing for Fairness Through Attribution-Based Data Augmentation", International Conference on Artificial Intelligence and Statistics (AISTATS) 2026. [arXiv] [code]
L. You ✉, L. Cao, M. Nilsson, B. Zhao, and L. Lei, "Distributional Counterfactual Explanation With Optimal Transport", International Conference on Artificial Intelligence and Statistics (AISTATS) 2025 (Oral, top 2%). [arXiv] [code]
EU-ABPLA: The European AI Based Personal Language Learning Assistant
EU-ABPLA: The European AI Based Personal Language Learning Assistant
This project aims to transform second language education by developing an AI tutor that provides personalized guidance through natural language processing, speech recognition, and grammar feedback.
The consortium consists of: Vestegnens Sprog- og Kompetencecenter (Denmark) DTU Engineering Technology (Denmark), Katholiek Onderwijs Vlaanderen (Belgium), Vages-grow (Slovakia), and Kurzy Zebra (Czech Republic).
Page updated
Google Sites
Report abuse