Personal profile
Hello there, I am currently an Associate Professor in Reasoning and Learning Research Group led by Yang Gao, at School of Intelligence Science and Technology at Nanjing University. Before this, I was a Postdoctoral Researcher at University of Alberta from 2021 to 2024. I received my Ph.D. in 2021 at Tianjin University, supervised by Jianye Hao. My PhD thesis is titled ‘Efficient Deep Multiagent Reinforcement Learning Based on Transfer Learning’.
My major research interests focus on deep Reinforcement Learning (RL), multiagent systems, and AI Agent, especially in building intelligent agents with complex reasoning ability, high efficiency, generalization, and scalability through transfer/multi-task learning, hierarchical learning, and opponent modeling. I am also interested in exploring continuous learning in open scenarios. Currently, I am working on 1) how to effectively and efficiently build AI agents upon generative AI and RL techniques; 2) how to effectively transfer knowledge in cross-domain settings; 3) how to improve MARL exploration considering the communication protocol and causal effect; 4) how to improve RL generalization and interpretability via symbolic planning/causal reasoning/program synthesis.
I am currently serving as a reviewer for JMLR, TPAMI, TNNLS, TMLR, MACH, JMLC, IEEE TCDS, and IEEE/CAA, and a member of the (senior) program committee (NeurIPS, AAAI, ICLR, IJCAI, ICML, AAMAS, UAI, ICRA, CoRL, CIKM, ECAI, DAI).
I am looking for research assistants/postdocs (master starting in Fall 2027/PhD students starting in Fall 2028), interested in (deep) reinforcement learning, multiagent learning, or AI Agents. If you are interested and have good programming skills and a reinforcement learning background, please email me your CV, transcripts, and future research proposal.
News
2026
- One paper (The Evolutionary Trajectory of Medical Multi-Agent Systems: A Survey) has been accepted by EMNLP 2026!
- Two papers (Thinking-Based Non-Thinking: Solving the Reward Hacking Problem in Training Hybrid Reasoning Models via Reinforcement Learning, MDTeamGPT: Mitigating Context Collapse and Enabling Self-Evolution in Medical Multi-Agent Reasoning) have been accepted by ACL 2026!
- One paper (SafeDialBench: A Fine-Grained Safety Evaluation Benchmark for Large Language Models in Multi-Turn Dialogues with Diverse Jailbreak Attacks) has been accepted by ICLR 2026!
2025
- Two papers (Faster Game Solving via Asymmetry of Step Sizes, Causality-Aware Efficient Exploration for Cooperative Multi-Agent Reinforcement Learning) have been accepted by AAAI 2026!
- Two papers (Parameter-Free Last-Iterate Convergence of Counterfactual Regret Minimization Algorithms, Multi-Agent Reinforcement Learning with Communication-Constrained Priors) have been accepted by NeurIPS 2025!
- Our paper (Reducing Variance of Stochastic Optimization for Approximating Nash Equilibria in Normal-Form Games) has been accepted by ICML 2025 as a Spotlight poster (top 2.6%)!
- One paper (The Evolving Landscape of LLM- and VLM-Integrated Reinforcement Learning) has been accepted by IJCAI 2025 Survey Track!
- Two papers (Towards Empowerment Gain through Causal Structure Learning in Model-Based RL, Causal Information Prioritization for Efficient Reinforcement Learning) have been accepted by ICLR 2025!
2024
- Two papers (Empowering Generalization for Deep Reinforcement Learning via Symbolic Planning, Taming Multi-Agent Reinforcement Learning with Estimator Variance Reduction) have been accepted by AAMAS 2025!
- I recently joined School of Intelligence Science and Technology at Nanjing University (Suzhou) as an Assistant Professor!
- I gave a talk at InterPol workshop 2024 at RLC 2024!
- I was invited as an area chair of AAMAS 2025!
- Our paper A survey on interpretable reinforcement learning has been published on Machine Learning!
- One paper (FPGA Divide-and-Conquer Placement using Deep Reinforcement Learning) has been accepted by ISEDA 2024!
- Our paper (LaFFi: Leveraging Hybrid Natural Language Feedback for Fine-tuning Language Models) received a Best Paper Runner-Up award from the HCRL@AAAI-24 workshop!
2023
- Two papers (PADDLE: Logic Program Guided Policy Reuse in Deep Reinforcement Learning, Mastering Robot Control through Point-based Reinforcement Learning with Pre-training) have been accepted by AAMAS 2024!
- Two papers (A Transfer Approach Using Graph Neural Networks in Deep Reinforcement Learning, PORTAL: Automatic Curricula Generation for Multiagent Reinforcement Learning) have been accepted by AAAI 2024! One paper (LaFFi: Leveraging Hybrid Natural Language Feedback for Fine-tuning Language Models) has been accepted by AAAI 2024 Workshop on Human-Centric Representation Learning!
- Our paper (Human-in-the-Loop Reinforcement Learning: A Survey and Position on Requirements, Challenges, and Opportunities) has been accepted by Journal of Artificial Intelligence Research (JAIR)!
- Two papers (PADDLE: Logic Program Guided Policy Reuse in Deep Reinforcement Learning, Work-in-Progress: Using Symbolic Planning with Deep RL) have been accepted by GenPlan Workshop NeurIPS 2023! One paper (Reinforcement Learning for FPGA Placement) has been accepted by MlSys Workshop NeurIPS 2023!
- Our paper (ASN: Action Semantics Network for Multiagent Reinforcement Learning) has been accepted by Autonomous Agents and Multi-Agent Systems (JAAMAS)!
- I was invited as an area chair of (MRS 2023)!
- Our paper (T3S: Improving Multi-Task Reinforcement Learning with Task-Specific Feature Selector and Scheduler) got accepted at IJCNN 2023!
- Our paper (Exploration in Deep Reinforcement Learning: From Single-Agent to Multi-Agent Domain) got accepted at IEEE Transactions on Neural Networks and Learning Systems (TNNLS)! Two papers (PORTAL: Automatic Curricula Generation for Multiagent Reinforcement Learning, Transfer Learning based Agent for Automated Negotiation) got accepted at AAMAS 2023 as Extended Abstract!
2022
- Our paper (Learning to Shape Rewards using a Game of Two Partners) got accepted at AAAI 2023! One paper (Efficient Deep Reinforcement Learning via Policy-extended Successor Feature Approximator) got accepted at DAI 2022!
- Our paper (GALOIS: Boosting Deep Reinforcement Learning via Generalizable Logic Synthesis) got accepted at NeurIPS 2022!
- Our paper (Cross-domain Adaptive Transfer Reinforcement Learning Based on State-Action Correspondence) got accepted at UAI 2022! One paper (PMIC: Improving Multi-Agent Reinforcement Learning with Progressive Mutual Information Collaboration) got accepted at ICML 2022!
Publications (To be updated)
A Transfer Approach Using Graph Neural Networks in Deep Reinforcement Learning. Tianpei Yang et al. AAAI. 2024. url
Portal: Automatic curricula generation for multiagent reinforcement learning. Jizhou Wu, Jianye Hao (Corresponding author), Tianpei Yang* (Corresponding author), Xiaotian Hao, Yan Zheng, Weixun Wang, Matthew E Taylor. AAAI. 2024. url
ASN: Action Semantics Network for Multiagent Reinforcement Learning. Tianpei Yang et al. JAAMAS. 2023. url
GALOIS: Boosting Deep Reinforcement Learning via Generalizable Logic Synthesis. Yushi Cao, Zhiming Li, Tianpei Yang* (Corresponding author) et al. NeurIPS. 2022. url
Cross-domain Adaptive Transfer Reinforcement Learning Based on State-Action Correspondence. Heng You, Tianpei Yang* (Corresponding author) et al. UAI. 2022.url
An Efficient Transfer Learning Framework for Multiagent Reinforcement Learning. Tianpei Yang et al. NeurIPS. 2021. url
Efficient Deep Reinforcement Learning via Adaptive Policy Transfer. Tianpei Yang et al. IJCAI. 2020. url
From Few to More: Large-scale Dynamic Multiagent Curriculum Learning. Weixun Wang (Equal contribution), Tianpei Yang (Equal contribution) et al. AAAI. 2020. url
Towards Efficient Detection and Optimal Response against Sophisticated Opponents. Tianpei Yang et al. IJCAI. 2019. url
To be updated