Scholar
Xiaochuan Li
Google Scholar ID: 97QHT-0AAAAJ
Carnegie Mellon University
Machine Learning
Natural Language Processing
Follow
Homepage
↗
Google Scholar
↗
Citations & Impact
All-time
Citations
540
H-index
5
i10-index
3
Publications
7
Co-authors
4
list available
Contact
CV
Open ↗
Twitter
Open ↗
GitHub
Open ↗
Publications
11 items
One Run Is Not an Idea: The Implementation Lottery in Automated Research
2026
Cited
0
ACM: Agentic Context Management for Long Horizon Tasks
2026
Cited
0
Auto Research for Materials: Auditable AI-Scientist Workflows with Held-Out Transfer
2026
Cited
0
Closed-loop Auto Research for Molecular Property Prediction: Discovering and Certifying Generalizable Improvements
2026
Cited
0
Auto Research with Specialist Agents Develops Effective and Non-Trivial Training Recipes
2026
Cited
0
Benchmark Test-Time Scaling of General LLM Agents
2026
Cited
0
Who Judges the Judge? LLM Jury-on-Demand: Building Trustworthy LLM Evaluation Systems
2025
Cited
0
Droplet3D: Commonsense Priors from Videos Facilitate 3D Generation
2025
Cited
0
Load more
Resume (English only)
Academic Achievements
Scaling Computer-Use Grounding via UI Decomposition and Synthesis, arXiv 2025 (*equal contribution)
Montessori-Instruct: Generate Influential Training Data Tailored for Student Learning, ICLR 2025
OSWorld: Benchmarking Multimodal Agents for Open-Ended Tasks in Real Computer Environments, NeurIPS 2024 D&B Track
Outstanding Graduate, Tsinghua University, 2025
Special-Class Scholarship, Tsinghua University, 2024
Comprehensive Excellence Scholarship, Tsinghua University, 2023
Background
First-year Ph.D. student at the Language Technologies Institute, Carnegie Mellon University
Research interests lie in Machine Learning and Natural Language Processing, with a particular focus on data-centric large language models
Interested in the relationship between data and model performance
Keen on tracing the origins of model capabilities (e.g., reasoning) to their data sources
Focused on generating higher-quality and more effective training data
Co-authors
4 total
Tao Yu
Assistant Professor, Computer Science, University of Hong Kong
Philip S. Yu
Professor of Computer Science, University of Illinons at Chicago
Xuming Hu
Assistant Professor, HKUST(GZ) / HKUST
Chenyan Xiong
Associate Professor, Carnegie Mellon University