Dr. Dawei Zhou is an Associate Professor at the Computer Science Department of Virginia Tech. Zhou’s primary research focuses on AI for Physical Intelligence and Scientific Discovery. His group develops intelligent systems that can autonomously generate, validate, and refine scientific hypotheses—advancing discovery pipelines in metamaterial design, healthcare, financial forensics, autonomous driving, and physics-guided predictive maintenance. He obtained his Ph.D. degree from the Computer Science Department of the University of Illinois Urbana-Champaign (UIUC). He has authored more than 60 publications in premier academic venues across AI, data mining, and information retrieval (e.g., ICML, NeurIPS, ICLR, KDD, WWW, TMLR) and has served as Vice Program Chair/Proceeding Chair/Student Travel Award Chair/Local Chair/ Session Chairs/(Senior) Program Committee Members in various top ML and AI conferences (e.g., KDD, WWW, ICDM, BigData, NeurIPS, ICML, ICLR, AAAI, IJCAI, etc.). VLOG lab received generous support from a diverse set of funding sources across federal agencies, industry, and private sectors, including NSF (CAREER, III, SCH), DARPA (FoundSci, Scify, Mathbac), DHS, the Commonwealth Cyber Initiative, 4-VA, Deloitte, Amazon, Google, and Cisco. His work has been recognized by the 24th CNSF Capitol Hill Science Exhibition, Cisco Faculty Research Award (2023), AAAI New Faculty Highlights Roster (2024), NSF Career Award (2024), National Distinction Program (2025), and Virginia Tech Outstanding Assistant Professor Award (2025).
Research Agenda: The VLOG Lab develops AI for physical intelligence and scientific discovery. Our goal is to build AI systems that can reason about physical systems, propose testable hypotheses and designs, and revise their conclusions as new evidence becomes available. We combine machine learning, scientific knowledge, and automated reasoning to help researchers decide what to test next, especially when data are scarce and experiments are costly.
Current Research:
Artificial Long-Tail Intelligence (NSF CAREER, etc.). We develop methods that learn from rare and underrepresented examples, recognize unfamiliar cases, and adapt as data distributions change. We study these problems in graph learning, with applications in areas such as rare-disease diagnosis and financial fraud detection.
Agentic AI for Scientific Discovery (DARPA SciFy, DARPA MathBac, etc.). How can AI agents work together to decide what to test next? We study how agents share evidence, coordinate long-horizon investigations, and choose among reasoning, simulation, and experiments based on what each step is expected to reveal.
Physics-Guided Materials Discovery and Design (DARPA FoundSci, etc.). We develop generative models and scientific agents that design mechanical metamaterials with target properties. Our work connects design generation, property prediction, simulation, and physics-based verification. Projects such as UniMate and MetamatBench support this research and provide testbeds for our DARPA-funded discovery workflows.
Reliable Reasoning and AI Assurance (NSF Future Core, etc.). We investigate when and why large language models hallucinate and how errors propagate across multistep reasoning. We develop methods to quantify uncertainty, check claims against evidence, and identify when an AI system needs further verification.
Agentic AI for Clinical Reasoning and Evidence Synthesis (NSF SCH, etc.). We study how AI agents can bring together patient histories, clinical guidelines, and biomedical literature to support clinical reasoning. Our research focuses on tracing conclusions back to their sources, identifying gaps or conflicts in the evidence, and determining which findings need clinician review.
Lab Openings: We welcome inquiries from prospective Ph.D. students and Postdoctoral Researchers interested in above areas. Applicants with backgrounds in computer science, statistics, physics, math, or engineering are encouraged to reach out.
Email zhoud@vt.edu with your CV, a brief description of your research interests, and the projects you would like to work on. Prospective Ph.D. students should also include their transcripts. Funded opportunities vary by project and start date.
[11/2026]: Honored to receive a new grant from DARPA Mathbac as VT PI to support our research on Agentic AI Communication.
[10/2026]: Honored to receive the ACM Computing Surveys (CSUR) 2025 Outstanding Associate Editor Award.
[9/2026]: Three papers have been accepted by NeurIPS2026. Congrats to Xinyue, Wangzhi, and Jianpeng, for their work on long-horizon reasoning and metamaterial design.
[8/2026]: On paper on Long-Tail-Aware Semantic ID has been accepted by ICDM2026. Congrats to Longfeng and Tong! This is a wonderful culmination of our Amazon Research Award.
[7/2026]: Two papers on Agentic Predictive Maintenance has been accepted by CIKM2026 and TMLR2026. Congrats to Shuaicheng and Tuo! This is a wonderful culmination of our four-year collaboration with Deloitte and VT NSI.
[6/2026]: One paper on Multi-Agent Language-Guided Metamaterial Discovery has been accepted by COLM 2026. Congrats to Jianpeng and Wangzhi!
[5/2026]: Two papers have been accepted by SIGKDD 2026. Congrats to Jianpeng and Wangzhi!
[5/2026]: Big congratulations to my student Longfeng Wu on passing her PhD final defense, "Towards the Next Generation of Recommendation Systems: Exploration, Trustworthiness, and Context-Aware Generation", and on her upcoming role as a AI Scientist at Amazon!
[5/2026]: Big congratulations to my student Shuaicheng Zhang on passing his PhD final defense, “Towards Open-World Graph Learning: Theory, Algorithm, and Deployment,” and on his upcoming role as a AI Scientist at LinkedIn!
[5/2026]: One paper on Physics-Grounded Material Claim Verification has been accepted by ACL 2026. Congrats to Jianpeng as the leading author!
[4/2026]: I am invited to deliver a Keynote Lecture on Open-World Long-Tailed Learning at the 65th Midwest Graph Theory Conference.
[4/2026]: Honored to receive a new grant from the Commonwealth Cyber Initiative as PI to support our research on LLM Hallucination Characterization.
[3/2026]: Honored to receive the Virginia Tech Big Contribution Grant. Congrats to Steve for his first leading PI grant since joining VT!
[2/2026]: I am delighted to serve as the Undergraduate and REU Consortium Chair for IEEE BigData 2026, which will be held in Phoenix, Arizona.
[1/2026]: Three papers have been accepted by ICLR 2026. Congrats to Xinyue and Junhong for their leading-author papers on LLM Hallucination Characterization and Test-Time Scaling for LLM Reasoning, respectively!
Office: 3160 F, Torgersen Hall, 620 Drillfield Dr., Blacksburg, VA 24060
Phone: (540) 231-2642
Email: zhoud [at] vt [dot] edu