Ping Liu
Assistant Professor, Department of Computer Science and Engineering, University of Nevada, Reno
Controllable, verifiable, and physically grounded generative AI.
About
Ping Liu is a Tenure-Track Assistant Professor in the Department of Computer Science and Engineering at the University of Nevada, Reno (UNR), where he has been since March 2024. From 2020 to 2024 he was a Scientist and then Senior Scientist II at the Centre for Frontier AI Research (CFAR), Institute of High Performance Computing (IHPC), A*STAR, Singapore. Before that, he was a Research Fellow and Lecturer at the University of Technology Sydney, a Research Scientist at the JD.com US Research Center, and a Senior Applied Research Engineer at the Sony US Research Center.
He received his Ph.D. in Computer Science and Engineering from the University of South Carolina, his M.Eng. from Huazhong University of Science and Technology, and his B.Eng. from Wuhan University of Technology. He is an IEEE Senior Member.
Research
His research develops methods that diagnose what a generative model has learned, intervene on it selectively, and verify the result. Current work follows three directions.
His research is supported by the National Science Foundation, including an NSF Major Research Instrumentation award for a large-scale GPU cluster for AI research at UNR. Read more about his research →
Collaboration and Openings
- He welcomes research collaborations with academic and industry partners.
- Prospective Ph.D. students are encouraged to apply to the UNR CSE Ph.D. program and mention his name.
News
- Oct 2026ReleaseThe project page and code for PhyMetric (NeurIPS 2026) are now available. PhyMetric diagnoses which physical laws text-to-video models violate, using 8,920 scene-grounded yes/no questions over 1,115 scenes in 50 physical categories. Project page Code
- Oct 2026TalkInvited talk at UC Merced: "Intervene and Verify: Toward Controllable Generative Models and World Models."
- Sep 2026Paper"PhyMetric: Diagnosing Physical Plausibility in Text-to-Video Generation via Scene-Level QA," led by Ph.D. student Siwei Meng, accepted to NeurIPS 2026, Evaluations and Datasets Track (corresponding author). Project page Code
- Sep 2026Paper"Calibrated Drift Control for Heterogeneous DPO" accepted to NeurIPS 2026 (first author).
- Sep 2026Paper"DOME: Drift-Adaptive On-Policy Motion Erasure in Video Diffusion Transformers" accepted to NeurIPS 2026 (first author).
- Aug 2026Paper"MUTE: Motion Concept Unlearning in Video Diffusion Models" accepted to ACM Multimedia 2026 (first author). arXiv
- Aug 2026Paper"When Physical Preferences Meet Semantic Constraints: Physical and Semantic Direct Preference Optimization for Text-to-Video Generation," led by Ph.D. student Siwei Meng, accepted to ACM Multimedia 2026 and selected for an oral presentation (corresponding author). arXiv
- Jul 2026ServiceCo-organizing the workshop "MIRAI Driving: Multimodal Intelligence and Reasoning for Autonomous and Intelligent Driving" at ACCV 2026, Osaka, Japan.
- Jun 2026ServiceInvited to serve as Senior PC Member for AAAI 2027.
- Jun 2026Paper"PhyMAGIC: Physical Motion-Aware Generative Inference with Confidence-Guided VLM," led by Ph.D. student Siwei Meng, accepted to ECCV 2026 (co-corresponding author). Project page
- Jun 2026Paper"RIGI: Rectifying Image-to-3D Generation Inconsistency via Uncertainty-aware Learning" accepted to IEEE Transactions on Image Processing (co-corresponding author).
- May 2026Paper"Where Concept Erasure Should Occur: Concept–Layer Alignment in Text-to-Video Diffusion Models" accepted to ICML 2026 (co-corresponding author).
- Apr 2026Paper"ABCDE: Agentic-Based Controlled Dynamic Erasure for Intent-Aware Safety Reasoning" accepted to Transactions on Machine Learning Research (first author).
- Mar 2026Preprint"PROBE: Diagnosing Residual Concept Capacity in Erased Text-to-Video Diffusion Models" released on arXiv (co-corresponding author). arXiv
- Feb 2026ServiceInvited to serve as Area Chair for ACM MM 2026.
- Feb 2026Paper"Closed-Form Concept Erasure via Double Projections" accepted to CVPR 2026 (corresponding author). arXiv
- Feb 2026ServiceWorkshop proposal "SynthIR: The First Workshop on Synthetic Content in Information Retrieval Ecosystems" accepted at SIGIR 2026.
- Jan 2026ServiceInvited to serve as Senior PC Member for IJCAI 2026.