Autonomous Agents and Multiagent Systems (AAMAS). Awarded annually to a single doctoral graduate worldwide for outstanding dissertation research in agent and multi-agent systems.
Spotlight paper, Conference on Neural Information Processing Systems, 2025
Oral presentation, AAAI Conference on Artificial Intelligence, 2025
AI Track, ACM Conference on Economics and Computation, 2024
Spotlight paper, Conference on Neural Information Processing Systems, 2022
Spotlight paper, International Conference on Learning Representations, 2022
International Conference on Learning Representations, 2021
Awarded to top 1% students at IIIS, Tsinghua University
Spotlight paper, Conference on Neural Information Processing Systems, 2020
Spotlight paper, International Conference on Learning Representations, 2020
Preprint Papers
LLM Advertisement based on Neuron Auctions
LLM Advertising2026NaiAD: Initiate Data-Driven Research for LLM Advertising
LLM Advertising2026How LLMs Are Persuaded: A Few Attention Heads, Rerouted
LLM2026The Memory Curse: How Expanded Recall Erodes Cooperative Intent in LLM Agents
LLM Agents2026Incentive-Aware Multi-Fidelity Optimization for Generative Advertising in Large Language Models
LLM Advertising2026LLM Active Alignment: A Nash Equilibrium Perspective
LLM Agents2026
Conference Papers
Duality for Optimal Multi-Item, Multi-Bidder Auction Design: Revenue Certificates through Deep Learning
Mechanism Design2026ACM ECOralPolicy-to-Language: Train LLMs to Explain Decisions with Flow-Matching Generated Rewards
2026ICLRPDFComposite Flow Matching for Reinforcement Learning with Shifted-Dynamics Data
2025BundleFlow: Deep Menus for Combinatorial Auctions by Diffusion-Based Optimization
2025NeurIPSPDFAdaptive Frontier Exploration on Graphs with Applications to Network-Based Disease Testing
2025NeurIPSPDFRobust Optimization with Diffusion Models for Green Security
2025UAIPDFOn Diffusion Models for Multi-Agent Partial Observability: Shared Attractors, Error Bounds, and Composite Flow
2025The Bandit Whisperer: Communication Learning for Restless Bandits
2025GemNet: Menu-Based, Strategy-Proof Multi-Bidder Auctions Through Deep Learning
2024Multi-Sender Persuasion: A Computational Perspective
2024ICMLPDFPosition: Social Environment Design Should be Further Developed for AI-based Policy-Making
2024ICMLPDFDeep Contract Design via Discontinuous Neural Networks
2023NeurIPSPDFSymmetry-Aware Robot Design with Structured Subgroups
2023ICMLPDFLow-Rank Modular Reinforcement Learning via Muscle Synergy
2022Non-Linear Coordination Graphs
2022Self-Organized Polynomial-Time Coordination Graphs
2022Context-Aware Sparse Deep Coordination Graphs
2022Celebrating Diversity in Shared Multi-Agent Reinforcement Learning
2021RODE: Learning Roles to Decompose Multi-Agent Tasks
2021DOP: Off-Policy Multi-Agent Decomposed Policy Gradients
2021Incorporating Pragmatic Reasoning Communication into Emergent Language
2020ROMA: Multi-Agent Reinforcement Learning with Emergent Roles
2020Influence-Based Multi-Agent Exploration
2020Learning Nearly Decomposable Value Functions with Communication Minimization
2020Convergence of Multi-Agent Learning with a Finite Step Size in General-Sum Games
2019AAMASPDFCompact Object Representation of a Non-Rigid Object for Real-Time Tracking in AR Systems
2018
Journal Papers
AM 220: Geometric Methods for Machine Learning
Spring 2026Teaching FellowDeep Reinforcement Learning
Spring 2020Teaching AssistantArtificial Intelligence: Principles and Techniques
Fall 2020Teaching AssistantArtificial Intelligence: Principles and Techniques
Fall 2019Teaching Assistant
International Conference on Learning Representations
2022–2026ICLRReviewerOutstanding Reviewer 2021Conference on Neural Information Processing Systems
2021–2026NeurIPSReviewerArea ChairInternational Conference on Machine Learning
2021–2026ICMLReviewerThe AAAI Conference on Artificial Intelligence
2020–2026AAAIReviewerInternational Joint Conference on Artificial Intelligence
2020–2026IJCAIReviewer
- Write: twang1@g.harvard.edu
- Drop in: F509, Zhongguancun Zhizao Street, Haidian, Beijing
Hi, I'm Tonghan Wang. I will be joining the Institute for AI at Tsinghua University as an Assistant Professor in 2026.
Research Interests
Current ongoing projects include:
- Advertising and market design for generative models, grounded in economic theory
- AI game theory with an emphasis on safety
- Embodied learning with coordinated muscle-group control
- Multi-agent problems in generative models
As the world moves toward human–agent coexistence, all questions about the relationship between humans and AI are within scope.
I am continuously looking for passionate PhD students and undergraduate research interns. I admit two PhD students per year; PhD admissions are currently open for students entering in 2027 or later (including the 2026 early-admission cycle). I encourage prospective students to first connect through a research internship or collaborative project — this also helps you assess whether my group is the right fit for you.
Undergraduate interns: Students at any stage of undergraduate study are welcome. We can tailor a research plan to your background and future goals.
If you are excited about our research directions, please send your CV and transcripts to twang1@g.harvard.edu, along with a brief description (one or two sentences) of a research problem you have worked on and what specific questions you find most interesting.
Mentoring Style
- Student development comes first. For students aiming for academic careers, I can facilitate collaborations or research visits with top institutions including Harvard, MIT, CMU, and Google.
- I encourage tackling hard and important problems, value originality and depth, and actively support collaboration across institutions and with industry.
- I have extensive personal experience navigating the complexities of U.S. visas 😂 and can offer detailed, practical advice.
About Me
- PhD from Harvard University, advised by Prof. David C. Parkes and Prof. Milind Tambe
- Master's from the Institute for Interdisciplinary Information Sciences (IIIS), Tsinghua University
- Victor Lesser Distinguished Dissertation Award, 2025 — awarded annually to a single doctoral graduate worldwide for outstanding dissertation research in agent and multi-agent systems
- Best Paper Award at ACM EC 2024 (AI Track); Oral at ACM EC 2026; multiple Spotlight/Oral recognitions at NeurIPS, ICLR, AAAI, and other top venues