Each member connects foundations, systems, interfaces, and applications around the PAI agenda.
PI
Lijie leads PAI Lab, connecting provable foundations, actionable systems, interaction, and evaluation for large-model interpretability.
Lijie leads PAI Lab, connecting provable foundations, actionable systems, interaction, and evaluation for large-model interpretability.
Postdoc
Wei is a postdoctoral researcher with research interests in machine learning, explainable AI, VLMs, and LLM agents.
Wei is a postdoctoral researcher with research interests in machine learning, explainable AI, VLMs, and LLM agents.
PhD Student
Jiahao studies interpretable AI for controllable generation, with interests in AI for biology, reasoning, and model control.
Jiahao studies interpretable AI for controllable generation, with interests in AI for biology, reasoning, and model control.
PhD Student
Hongji studies multimodal recommendation, user behavior modeling, value prediction, LLM alignment, and agent coordination.
Hongji studies multimodal recommendation, user behavior modeling, value prediction, LLM alignment, and agent coordination.
PhD Student
Junchi explores LLMs, multimodal agents, world models, VLA systems, and model-behavior evaluation.
Junchi explores LLMs, multimodal agents, world models, VLA systems, and model-behavior evaluation.
Master Student
Zeqing builds AI-powered applications and full-stack systems, with interests in NLP, speech, data analysis, and applied machine learning.
Zeqing builds AI-powered applications and full-stack systems, with interests in NLP, speech, data analysis, and applied machine learning.
Master Student
Daniil received his BSc from CUHK and is interested in LLM interpretability and reinforcement learning.
Daniil received his BSc from CUHK and is interested in LLM interpretability and reinforcement learning.
Visiting Student
Zikang studies LLM-agent security, adversarial manipulation, privacy leakage, and behavioral risks in multi-step decision-making.
Zikang studies LLM-agent security, adversarial manipulation, privacy leakage, and behavioral risks in multi-step decision-making.
Visiting Student
Tingxu is a visiting researcher from Nanjing University. His research interests include Responsible AI, LLM Reasoning, and AI4SE.
Tingxu is a visiting researcher from Nanjing University. His research interests include Responsible AI, LLM Reasoning, and AI4SE.
Visiting Student
Ziyue‘s research interests span the broad fields of machine learning and neural networks, with a particular focus on generative AI, multi-modal understanding, explainability, neuro-symbolic and continual learning
Ziyue‘s research interests span the broad fields of machine learning and neural networks, with a particular focus on generative AI, multi-modal understanding, explainability, neuro-symbolic and continual learning
Visiting Student
Luqing studies distributed optimization, federated learning, control theory, differential privacy, LLM unlearning/security, and representation analysis.
Luqing studies distributed optimization, federated learning, control theory, differential privacy, LLM unlearning/security, and representation analysis.
Research Assistant
Peng studies contrastive learning, representation analysis, and LLM methods, with recent ICML work on Bayesian gating.
Peng studies contrastive learning, representation analysis, and LLM methods, with recent ICML work on Bayesian gating.
