Lei Han (韩磊)

Senior Research Scientist

Tencent AI Lab
Tencent, Shenzhen, China

Office: Lang Ke Building, Shenzhen, China.
Tencent Email: lxhan at tencent dot edu
Gmail: leihan dot cs at gmail dot com

Previous email addresses: lei.han@msstate.edu; lhan@stat.rutgers.edu; leihan@comp.hkbu.edu.hk; hanlei@cis.pku.edu.cn


I am currently a senior research scientist in Tencent AI Lab. I was an assistant research professor in the Department of Basic Science, Mississippi State University, USA. I received my Ph.D degree from Peking University (advised by Professor Kunqing Xie) and spent two years in Hong Kong Baptist University (advised by Professor Yu Zhang) and Rutgers University (advised by Professor Tong Zhang) as postdoctoral researcher. I was the winner of the Best Dissertation Award of Chinese Association for Artificial Intelligence (CAAI) (中国人工智能学会优博).

My research interests mainly focus on machine learning and artifical intelligence. I am especially interested in large-scale statistical machine learning, reinforcement learning, optimization, multi-task learning and their applications in game AI, NLP, CV and bioinformatics.

Internship positions are available now. Please feel free to send me an email if you are interested.

News

  • [Nov. 2019] One paper have been accepted by AAAI 2020.
  • [Sep. 2019] Two papers have been accepted by NeurIPS 2019.
  • [May 2019] The Local Uncertainty Sampling paper has been accepted by Annals of Statistics.
  • [Apr. 2019] Our GridNet paper has been accepted by ICML 2019. Congrats to the team!
  • [Sep. 2018] Our AI agents are able to defeat the cheating level builtin AI in StarCraft II full game.

Selected Publications

[Google Scholar]

Preprints

  • Qing Wang, Jiechao Xiong, Lei Han, Meng Fang, Xinghai Sun, Zhuobin Zheng, Peng Sun, Zhengyou Zhang. (* equal contribution)
    Arena: a toolkit for Multi-Agent Reinforcement Learning. [arxiv]
    arXiv:1907.09467 [cs.LG], 2019.
  • Yiheng Huang, Liqiang He*, Guangsen Wang, Lei Han and Dan Su. (* equal contribution)
    Phrase-Level Class based Language Model for Mandarin Smart Speaker Query Recognition. [arxiv]
    arXiv:1909.00556 [cs.CL], 2019.
  • Jiechao Xiong, Qing Wang, Zhuoran Yang, Peng Sun, Lei Han, Yang Zheng, Haobo Fu, Tong Zhang, Ji Liu, Han Liu
    Parametrized Deep Q-Networks Learning: Reinforcement Learning with Discrete-Continuous Hybrid Action Space. [arxiv]
    arXiv:1810.06394 [cs.LG], 2018.

  • Peng Sun*, Xinghai Sun*, Lei Han*, Jiechao Xiong*, Qing Wang, Bo Li, Yang Zheng, Ji Liu, Yongsheng Liu, Han Liu, Tong Zhang. (*equal contribution)
    TStarBots: Defeating the Cheating Level Builtin AI in StarCraft II in the Full Game. [arxiv]
    arXiv:1809.07193 [cs.AI], 2018.

2020

  • [28] Xiangtai Li, Houlong Zhao, Lei Han, Yunhai Tong, Shaohua Tan, Kuiyuan Yang.
    Gated Fully Fusion for Semantic Segmentation.
    In: The Thirty-Fourth AAAI Conference on Artificial Intelligence (AAAI), 2020.

2019

  • [27] Yali Du*, Lei Han*, Meng Fang, Ji Liu, Tianhong Dai, Dacheng Tao. (* equal contribution)
    LIIR: Learning Individual Intrinsic Reward in Multi-Agent Reinforcement Learning.
    In: The Thirty-third Annual Conference on Neural Information Processing Systems (NeurIPS), 2019.
  • [26] Meng Fang, Tianyi Zhou, Yali Du, Lei Han, Zhengyou Zhang.
    Curriculum-guided Hindsight Experience Replay.
    In: The Thirty-third Annual Conference on Neural Information Processing Systems (NeurIPS), 2019.
  • [25] Lei Han, Kean Ming Tan, Ting Yang and Tong Zhang.
    Local Uncertainty Sampling for Large-Scale Multi-Class Logistic Regression. [arxiv]
    In: Annals of Statistics (AOS), 2019.
  • [24] Lei Han*, Peng Sun*, Yali Du*, Jiechao Xiong, Qing Wang, Xinghai Sun, Han Liu, Tong Zhang. (* equal contribution)
    Grid-Wise Control for Multi-Agent Reinforcement Learning in Video Game AI. [PDF][Supplementary Material][demo video]
    In: The Thirty-sixth International Conference on Machine Learning (ICML), 2019.
  • [23] Yu Zhang and Lei Han.
    Learning (from) Deep Hierarchical Structure among Features. [PDF][Link]
    In: The Thirty-Third AAAI Conference on Artificial Intelligence (AAAI), 2019.

2018

  • [22] Qing Wang, Jiechao Xiong, Lei Han, Peng Sun, Han Liu and Tong Zhang.
    Exponentially Weighted Imitation Learning for Batched Historical Data. [PDF][Link]
    In: The Thirty-second Annual Conference on Neural Information Processing Systems (NeurIPS), 2018.

  • [21] Lei Han, Yiheng Huang and Tong Zhang.
    Candidates vs Noises Estimation for Large Multi-Class Classification Problem. [PDF][Link][arxiv]
    In: The 35th International Conference on Machine Learning (ICML), 2018. (Long Presentation)

  • [20] Lei Han, Lei Li, Feng Wen, Lei Zhong, Tong Zhang and Xiu-Feng Wan.
    Graph-Guided Multi-Task Sparse Learning Model: a Method for Identifying Antigenic Variants of Influenza A(H3N2) Virus.
    In: Bioinformatics, 2018.

  • [19] Dong Dai, Lei Han, Ting Yang and Tong Zhang.
    Bayesian Model Averaging with Exponentiated Least Squares Loss. [Link][arxiv]
    In: IEEE Transactions on Information Theory (TIT), 2018.

  • [18] Sichen Du, Guojie Song, Lei Han and Haikun Hong.
    Temporal Causal Inference with Time Lag. [Link]
    In: Neural Computation 30 (1), 271-291, 2018.

2016

  • [17] Lei Han, Yu Zhang, Xiu-Feng Wan and Tong Zhang.
    Generalized Hierarchical Sparse Model for Arbitrary-Order Interactive Antigenic Sites Identification in Flu Virus Data. [PDF][Link][Supplementary Material][Code]
    In: Proceedings of the 22nd ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD), San Francisco, USA, 2016. (Acceptance Rate = 18.1%. Full presentation with acceptance rate = 8.9%)

  • [16] Lei Han*, Yu Zhang* and Tong Zhang (*equal contribution)
    Fast Component Pursuit for Large-Scale Inverse Covariance Estimation. [PDF][Link]
    In: Proceedings of the 22nd ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD), San Francisco, USA, 2016. (Acceptance Rate = 18.1%)

  • [15] Lei Han and Yu Zhang. (Both authors contributed equally)
    Reduction Techniques for Graph-based Convex Clustering. [PDF][Link][Supplementary Material]
    In: Proceedings of the Thirtieth AAAI Conference on Artificial Intelligence (AAAI), Phoenix, Arizona, USA, 2016. (Acceptance Rate = 26%)

  • [14] Lei Han and Yu Zhang. (Both authors contributed equally)
    Multi-Stage Multi-Task Learning with Reduced Rank. [PDF][Link][Supplementary Material]
    In: Proceedings of the Thirtieth AAAI Conference on Artificial Intelligence (AAAI), Phoenix, Arizona, USA, 2016. (Acceptance Rate = 26%)

  • [13] Lei Li, Lei Han and Xiu-Feng Wan.
    Identification of glycosylation sites and mutations determining antigenic drift events for influenza A viruses using sparse group lasso regression. [Link]
    In: GLYCOBIOLOGY 26 (12), 1393-1394, 2016.

  • [12] Xiabing Zhou, Xingxing Xing, Lei Han, Haikun Hong, Kaigui Bian, Kunqing Xie.
    Structure feature learning method for incomplete data. [PDF][Link][Supplementary Material]
    In: International Journal of Pattern Recognition and Artificial Intelligence 30 (9): 1660007, 2016.

2015

  • [11] Lei Han and Yu Zhang. (Both authors contributed equally)
    Learning Tree Structure in Multi-Task Learning. [PDF][Link][Supplementary Material][Code]
    In: Proceedings of the 21st ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD), Sydney, 2015. (Acceptance Rate = 19%)

  • [10] Xiabing Zhou, Lei Han, Xingxing Xing, Haikun Hong, Wenhao Huang, Kaigui Bian and Kunqing Xie.
    Incorporating temporal smoothness and group structure in learning with incomplete data. [Link]
    In: Proceedings of 12th International Conference on Fuzzy Systems and Knowledge Discovery (FSKD), 2015.

  • [9] Ye Liu, Liqiang Nie, Lei Han, Luming Zhang, David Rosenblum.
    Action2Activity: Recognizing Complex Activities from Sensor Data. [Link]
    In: International Joint Conference on Artificial Intelligence (IJCAI), 2015. (Acceptance Rate = 28.8%)

  • [8] Guojie Song, Lei Han* and Kunqing Xie. (* The corresponding author; the first two authors contributed equally)
    Overlapping Decomposition for Gaussian Graphical Modeling. [PDF][Link]
    In: IEEE Transactions on Knowledge and Data Engineering (TKDE), 2015. (An improved version of the conference paper appeared in KDD2012)

  • [7] Lei Han and Yu Zhang. (Both authors contributed equally)
    Learning Multi-Level Task Groups in Multi-Task Learning. [PDF][Link][Supplementary Material][Code]
    In: Proceedings of the Twenty-Ninth AAAI Conference on Artificial Intelligence (AAAI), Austin Texas, USA, 2015. (Acceptance Rate = 26.7%)

  • [6] Lei Han and Yu Zhang. (Both authors contributed equally)
    Discriminative Feature Grouping. [PDF][Link][Code]
    In: Proceedings of the Twenty-Ninth AAAI Conference on Artificial Intelligence (AAAI), Austin Texas, USA, 2015. (Acceptance Rate = 26.7%)

Before 2014

  • [5] Lei Han, Yu Zhang, Guojie Song and Kunqing Xie.
    Encoding tree-sparsity in Multi-Task Learning: A Probabilistic Framework. [PDF][Link]
    In: Proceedings of the Twenty-Eighth AAAI Conference on Artificial Intelligence (AAAI), Quebec City, Quebec, Canada, 2014. (Acceptance Rate = 28%)

  • [4] Lei Han, Guojie Song, Gao Cong and Kunqing Xie.
    Overlapping Decomposition for Causal Graphical Modeling. [PDF][Link]
    In: Proceedings of the 18th ACM SIGKDD international conference on Knowledge Discovery and Data Mining (KDD), Beijing, China, 2012. (Acceptance Rate = 18%)

  • [3] Lei Han, Kunqing Xie and Guojie Song.
    Adaptive Fit Parameters Tuning with Data Density Changes in Locally Weighted Learning. [PDF]
    In: Proceedings of the 7th International Symposium on Neural Networks (ISNN), Shanghai, China, 2010.

  • [2] Lei Han, Jianying Wu, Ping Gu, Kunqing Xie, Guojie Song, Shiwei Tang, Dongqing Yang, Bingli Jiao and Feng Gao.
    Adaptive Knowledge Transfer based on Locally Weighted Learning. [PDF]
    In: Proceedings of the Conference on Technologies and Applications of Artificial Intelligence (TAAI), Hsinchu, Taiwan, 2010.

  • [1] Lei Han, Meng Shuai, Kunqing Xie, Guojie Song and Xiujun Ma.
    Locally Kernel Regression Adapting with Data Distribution in Prediction of Traffic Flow. [PDF]
    In: Proceedings of the 18th International Conference on Geoinformatics (Geoinformatics), Beijing, China, 2010.

Ph.D. Thesis (In Chinese)

  • Lei Han. Traffic Network based Multi-Task Learning Method. EECS, Peking University, July, 2014.

Activities

  • Journal Reviewer: IEEE Transactions on Knowledge and Data Engineering (TKDE)
    IEEE Transactions on Intelligent Transportation Systems (TITS)
    Journal of Machine Learning Research (JMLR)

  • PC Member / Reviewer: AAAI 2016, NIPS 2016, AAAI 2017, NIPS 2017, AAAI 2018, ICLR 2018, ICML 2018, NIPS 2018, ICML 2019

Honors and Awards

  • The Best Dissertation Award of Chinese Association for Artificial Intelligence (CAAI) (中国人工智能学会优博), 2016. [Link]

  • Outstanding Ph.D. Graduate Award in Beijing, 2014

  • Outstanding Ph.D. Graduate Award in Peking University, 2014

  • President Scholarship (the highest Scholarship in Peking University), 2013-2014

  • President Scholarship (the highest Scholarship in Peking University), 2012-2013

  • Chinese National Scholarship (selected from top Ph.D. students in China), 2011-2012

  • President Scholarship (the highest Scholarship in Peking University), 2011-2012

  • President Scholarship (the highest Scholarship in Peking University), 2010-2011

  • Merit Student Award (selected from top postgraduates in Peking University), 2010-2011

  • President Scholarship (the highest Scholarship in Peking University), 2009-2010

  • Merit Student Award (selected from top postgraduates in Peking University), 2009-2010