My name is Wei Chen (陈 卫). I am a Senior Researcher at Microsoft Research Asia. I am also an Adjunct Professor in the Institute of Interdisciplinary Information Sciences, Tsinghua University and an Adjunct Researcher in the Institute of Computing Technology, Chinese Academy of Sciences.
My main research interests include social and information networks, online learning, algorithmic game theory, Internet economics, distributed computing, and fault tolerance. My research work aims at connecting real-world applications and phenomena with theoretical foundations through mathematical and algorithmic tools. I have been serving on the PCs or Senior PCs of the top data mining, database, machine learning and artificial intelligence conferences, such as KDD, WSDM, WWW, SDM, SIGMOD, ICDE, NIPS, ICML, IJCAI, AAAI, etc. I am a member of Technical Committees of Big Data and Theoretical Computer Science of Chinese Computer Federation.
I obtained my bachelor and master degrees from the Department of Computer Science and Technology, Tsinghua University, and my Ph.D degree from the Department of Computer Science, Cornell University. I worked for Oracle Corporation before joining MSRA.
I served as the captain of Tsinghua University Varsity soccer team, which won Beijing Inter-collegiate Championship twice in 1992 and 1993. I also served as the captain of the “Tsinghua Veterans” soccer team in North America, which won the championship title of the Annual North America Chinese Soccer Tournament twice in 1996 and 1997. I am still an active member of Tsinghua Forerunners (领跑) alumni soccer team, which won the Tsinghua Alumni Cup in 2012 and Tsinghua Alumni Premier League in 2018.
Professional association committees
- Member of Task Force on Big Data, China Computer Federation (CCF), 2012 – present
- Member of Technical Committee on Theoretical Computer Science, China Computer Federation (CCF), 2017 – present
- Organizer of the International Workshop on Computational Aspects of Social and Information Networks (CASIN’2011), July 20-22, 2011, Beijing, China.
- Organizer of Microsoft Research Asia Theory Workshop, April 2008
Editorial boards/Journal guest editors
- Associate Editor of ACM Transactions on Intelligent Systems and Technology (TIST).
- Member of the editorial board of Big Data Research, China Posts and Telecom Press.
- Member of the associate editorial board of Computational Social Networks, SpringerOpen.
- Guest co-editor for the special issue on Computational Aspects of Social and Information Networks (CASIN) at ACM Transactions on Knowledge Discovery in Data (CASIN-TKDD)
Conference program committees, journal reviewers
- PC co-chair for the 11th ACM/IEEE International Conference on Advances in Social Networks Analysis and Mining (ASONAM), 2019
- PC co-chair for the 26th International World Wide Web Conference (WWW’17) Poster Track, 2017.
- Program committee members for major AI/ML conferences:
- Annual Conference on Neural Information Processing Systems (NeurIPS/NIPS), 2016 – 2019
- International Joint Conference on Artificial Intelligence (IJCAI), 2019 (senior member)
- AAAI Conference on Artificial Intelligence (AAAI), 2019
- International Conference on Machine Learning (ICML), 2018
- Program committee members for major data mining / data management / big data conferences:
- ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD), 2013 – 2019
- International Conference on Web Search and Data Mining (WSDM), 2013, 2015 – 2016 (senior member)
- SIAM Data Mining conference (SDM), 2015 – 2017
- European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases (ECMLPKDD), 2015
- IEEE International Conference on Data Engineering (ICDE), 2014
- IEEE International Conference on Big Data (BigData), 2013, 2016-2017, 2018 (senior member)
- International Conference on Advanced Data Mining and Applications (ADMA), 2011 (vice co-chair), 2012
- Program committee members for major Web and social network conferences
- International World Wide Web Conference (WWW/WebConf), 2013 – 2019
- Social Influence Analysis Workshop (SocInf’15) at IJCAI, 2015
- IEEE/ACM International Conference on Advances in Social Networks Analysis and Mining (ASONAM), 2015
- International AAAI Conference on Weblogs and Social Media (ICWSM), 2012
- Program committee members for theoretical computer science conferences
- Ad Auction Workshop (AdAuction), 2015
- Annual International Computing and Combinatorics Conference (COCOON), 2013
- Annual Conference on Theory and Applications of Models of Computation (TAMC), 2010, 2012 – 2013
- International Frontiers of Algorithmics Workshop (FAW), 2009, 2011
- International Symposium on Algorithms and Computation (ISAAC), 2009
- International Conference on Algorithmic Aspects in Information and Management (AAIM), 2008, 2011
- Program committee members for distributed computing and fault tolerance conferences
- International Colloquium on Structural Information and Communication Complexity (SIROCCO), 2011, 2013
- IEEE International Conference on Distributed Computing Systems (ICDCS), 2005, 2008 – 2010
- International Conference On Distributed Computing and Networking (ICDCN), 2011
- International Conference On Principle Of Distributed Systems (OPODIS), 2009 – 2010
- Annual ACM SIGACT-SIGOPS Symposium on Principles of Distributed Computing (PODC), 2010
- Annual IEEE/IFIP International Conference on Dependable Systems and Networks (DSN), 2006, 2009
- International Symposium on Stabilization, Safety, and Security of Distributed Systems (SSS), 2008
- International Symposium on Reliable Distributed Systems (SRDS), 2007 – 2008
- Pacific Rim International Symposium on Dependable Computing (PRDC), 2005 – 2006
- Ad hoc referee for Nature, ACM Transactions on Knowledge Discovery from Data, ACM Transactions on Computation Theory, Journal of Machine Learning Research, ACM Transactions on the Web, ACM Transactions on Intelligent Systems and Technology, IEEE Transactions on Knowledge and Data Engineering, Distributed Computing, SIAM Journal on Computing, IEEE Transactions on Parallel and Distributed Systems, IEEE Transactions on Software Engineering, Journal of Parallel and Distributed Computing, IEEE International Conference on Dependable Systems and Networks, IEEE International Symposium on Reliable and Distributed Systems, International Symposium on Distribute Computing, Information Processing Letters, Parallel Processing Letters, Journal of Combinatorial Optimization, Mathematics of Operations Research, etc.
- Expert reviewers for National Science Foundation of China
Talks and Tutorials
Tutorial and Guest Lectures
- Guest lecture at Peking University: Information diffusion in social networks. Nov. 18, 2015. Also given at Tsinghua University.
- Guest lecture at Tsinghua University: Online learning. June 9, 2015.
- KDD 2012 Tutorial: Infomation and influence spread in social networks, with Carlos Castillo and Laks V. S. Lakshamanan. Beijing, China, Aug. 12, 2012.
Research Keynotes and Talks
- NCTCS’2017 (中国理论计算机年会) Keynote Speech: Combinatorial Online Learning. Wuhan, China, Oct. 14, 2017
- WWW’2017 talk: Interplay between social influence and network centrality: A comparative study of Shapley centrality and single-node-influence centrality. Perth, Australia, April 6, 2017
- Chinese National Computer Conference (CNCC) 2016 invited talk in the Online Algorithm Forum: Combinatorial Online Learning. Taiyuan, China, October 21, 2016.
- KDD’2016 talk: Robust influence maximization. San Francisco, U.S.A., August 15, 2016.
- Keynote speech at IJCAI’15 Social Influence Analysis Workshop: Computational Social Influence. Buenos Aires, Argentina, July 27, 2015.
- WWW’2015 talk: A game theoretic model for the formation of navigable small-world networks. Florence, Italy, May 21, 2015.
- Talk at CS Department, University of British Columbia: Combinatorial learning for combinatorial optimization — A trilogy. Vancouver, Canada, March 27, 2015.
- NIPS’2014 Full Oral Presentation: Combinatorial pure exploration in multi-armed bandits. Montreal, Canada, Dec. 10, 2014.
- Invited talk at the Institute of Mathematics and System Sciences, Chinese Academy of Sciences:Influence maximization: The new frontier — non-submodular optimizations. Beijing, China, Sept. 15, 2014.
- Keynote speech at the 10th Chinese National Mathematical Programming Conference: Influence diffusion modeling and optimizations in social networks. Luoyang, China, May 11, 2014.
- ACM EC’2013 talk: Sybil-proof mechanisms in query incentive networks. Atlanta, U.S.A., June 20, 2013.
- ICML’2013 talk (revised): Combinatorial multi-armed bandit: General framework, results, and applications. Philadelphia, U.S.A., June 18, 2013.
- Chinese National Computer Conference (CNCC) 2012 invited talk: Influence diffusion dynamics and influence maximization in complex social networks. Dalian, China, Oct. 18, 2012.
- Talk at EconCS Group of Harvard University.: Influence diffusion dynamics and influence maximization in complex social networks. Cambridge, MA. U.S.A, Oct. 18, 2011.
- CASIN 2011 talk: Influence maximization when negative opinions may emerge and propagate. Beijing, China, July 20, 2011.
- ICDM 2010 talk: Scalable influence maximization in social networks under the linear threshold model. Sydney, Australia, Dec. 15, 2010.
- KDD 2010 talk: Scalable influence maximization for prevalent viral marketing in large-scale social networks. Washington D.C, U.S.A., July 27, 2010.
Influence diffusion dynamics and influence maximization in social networks
– Monograph summarizing research results on modeling, optimization, and learning of information and influence diffusions: “Information and Influence Propagation in Social Networks, Morgan and Claypool, 2013“.
– Multi-round influence maximization [KDD’18]: We study the multi-round influence model where propagation from potentially different seeds occurs independently in different rounds, with the goal to collectively active nodes in social networks. This corresponds to multiple stages in a marketing campaign. We develop scalable algorithm for both non-adaptive and adaptive settings.
– Influence-based centrality [WWW’17]: We propose the study of influence-based network centrality and provide a comparative study on two centrality measures: single node influence (SNI) centrality and Shapley centrality. We provide axiomatic characterizations for both centralities as well as scalable centrality computation algorithms based on the reverse influence sampling approach.
– Robust influence maximization [KDD’16, arXiv:1601.06551]: We address the issue of robust optimization when parameters for the influence maximization are not accurate. We propose a new algorithm to achieve robust influence maximization with a guarantee, and investigate sampling methods to effectively improve parameter accuracy for the robust influence maximization task.
– Influence diffusion model covering the spectrum from competition to complementarity [VLDB’16, arXiv:1507.00317]: We propose an extension to the classical independent cascade model to cover the entire spectrum from competition to complementarity for two propagating items in a social network. We study the properties of the model and two variants of optimization problems for the mutually complementary regime. The paper also proposes a general sandwich approximation and a general condition for applying reverse reachable set approach to any diffusion model.
– Amphibious influence maximization [EC’15, arXiv:1507.03328]: we propose the model of combining traditional media marketing with viral marketing by selecting seed content providers to initiate cascades and selecting seed consumers to propagate the cascades in the social network. We prove hard-to-approximate result of the problem for general graphs, and propose an approximation algorithm for a certain class of restricted bipartite graph between content providers and consumers.
– Seed minimization with probabilistic coverage guarantee [KDD’14, arXiv: 1402.5516]: We look into the optimization problem of minimizing the seed set to ensure a probabilistic influence coverage guarantee. The main difficulty is that the objective function is not submodular. We provide theoretical analysis on the approximation guarantee of this problem and experimental study on the effectiveness of our algorithm.
– Active friending in social networks [KDD’13, arXiv: 1302.7025]: Active friending provides a new perspective on using social influence in social networks. Instead of finding seed nodes to maximize influence, it needs to find influence pathways to maximize influence. We provide efficient algorithms for finding intermediate nodes so as to increase the probability of a target node accepting the friending request.
– Scalable influence maximization: We study new algorithms and heuristics to improve the speed of finding influential nodes in a social network for viral marketing.
- MixedGreedy and DegreeDiscount [KDD’09]: improve the greedy algorithm by edge sampling, and propose DegreeDiscount for uniform independent cascade model.
- PMIA [KDD’10, DAMI’12]: scalable heuristic for the independent cascade model, using local tree structures to speed up influence computation for 3 orders of magnitude.
- LDAG [ICDM’10, MSR-TR-2010-133]: scalable heuristic for the linear threshold model, using local directed acyclic graph structures to speed up influence computation for 3 orders of magnitude.
- IRIE [ICDM’12, arXiv:1111.4795]: even faster scalable heuristic combining efficient influence ranking with influence estimation method, achieving up to 2 orders of magintude faster than PMIA in the independent cascade model.
- Dataset used: [data for two collaboration graphs NetHEPT and NetPHY][DBLP graph data]
- Code is available upon request by email
– Influence diffusion modeling and maximization for complex social interactions: We model complex social interactions including negative opinions due to product defects, competitive influence diffusion, and friend/foe relationships, and study influence maximization problem in these contexts.
- IC-N model and MIA-N algorithm [SDM’11, MSR-TR-2010-137]: We extend the independent cascade model to incorporate the emergence and propagation of negative opinions due to product defects, and study the influence maximization problem in this new model.
- Competitive influence diffusion CLT model and influence blocking maximization algorithm CLDAG in [SDM’12, arXiv:1110.4723]: We study the competitive linear threshold model, and propose efficient heuristic algorithm CLDAG for selecting the positive seed set that most effectively block the diffusion of negative influence.
- Time-critical influence maximization with time-delayed diffusion process [AAAI’12, arXiv:1204.3074]: We study influence maximization in which diffusion on each step may be delayed, and the objective is to maximize influence spread within a certain deadline. Both IC and LT models are extended, and efficient algorithms are proposed and evaluated.
- Influence diffusion dynamics and maximization in networks with friend and foe relationships [WSDM’13, arXiv:1111.4729]: We extend the voter model to signed networks representing both friend and foe relationships, and study the influence diffusion dynamics and influence maximization problem in this context.
– Participation maximization based on influence in online discussion forums [ICWSM’11, MSR-TR-2010-142]: We propose and study the participation maximization problem in online discussion forums, in which forum participation is determined by influence propagation through forum users.
Combinatorial online learning
– Combinatorial multi-armed bandit [ICML’13, JMLR’16, NIPS’17]: We provide a general stochastic framework that encompasses a large class of combinatorial multi-armed bandit problems, including many non-linear reward problems not considered before. We provide a CUCB learning algorithm with a tight analysis to show its regret bound. The framework can be applied to solve online learning problems with nonlinear reward functions such as probabilistic maximum coverage for online advertising, social influence maximization, as well as many optimization problems with linear reward functions. We further extend the model to cover probabilistically triggered arms, with applications to online influence maximization and combinatorial cascading bandits. With the latest improvement, we show that our algorithm could achieve good regret bounds for a large class of problems satisfying triggering probability modulated bounded smoothness conditions.
– Thompson sampling for CMAB [ICML’18]: We apply the Thompson sampling approach to combinatorial semi-bandit problems and provide theoretical regret bounds and proper conditions for the bounds to hold.
– CMAB with general reward functions [NIPS’16]: We address the combinatorial semi-bandit cases where the reward function not only depends on the means of the base arms, but on the entire distributions of the base arms. We design stochastic dominant confidence bound (SDCB) algorithm and show its regret bound, and apply it to K-Max problem and expected utility maximization problems.
– Contextual combinatorial cascading bandit [ICML’16]: We incorporate contextual information, together with position discounts and general reward functions to the combinatorial cascade bandit, which has applications in online advertising and online recommendations.
– Online greedy learning [NIPS’15]: We show how to turn an offline greedy algorithm into an online greedy learning algorithm, with competitive regret bound compared with the solution of the offline greedy algorithm.
– Combinatorial pure exploration [NIPS’14]: We study the problem of how to explore stochastic arms efficiently so that in the end we can output a subset of arms satisfying certain combinatorial constraint (k-set, matching, spanning tree, etc.) and maximizing the sum of expected rewards of these arms with high probability. We provide generic fix-confidence and fix-budget algorithms and prove a lower bound on sample complexity, which implies that our fix-confidence algorithm is tight (up to a logarithmic factor) for combinatorial constraints formed by bases of a matroid.
– Combinatorial partial monitoring [ICML’14, supplementary material]: We study the online learning problem where the optimization problem is combinatorial (e.g. matching) and thus the action space is exponentially large, and the feedback is limited. We provide efficient solutions to this problem, and demonstrate how to apply it to problems such as crowd sourcing task allocations.
Properties of small-world networks
– Navigability of small-world networks [WWW’15, arXiv:1411.4097]: We provide a game-theoretic formulation of navigable small-world networks to explain why real networks are often navigable. We provide strong theoretical and empirical results showing that navigable small-world network is the only stable equilibrium in the game. Our payoff function balances distance with relationship reciprocity, and this is the first work providing a surprising connection between relationship reciprocity and network navigability.
– Complex routing in small-world networks [COCOON’16, arXiv:1503.00448]: We define the decentralized routing problem of the complex contagion, which is the diffusion behavior in which a node is activated if and only if the number of its active friends is above a certain threshold greater than 1. We study the routing efficiency of complex contagions, and show that in Kleinberg’s small-world networks, the decentralized routing time is polynomial to the size of the graph in all parameter range, which means it cannot be efficient.
– The hyperbolicity of small-world and tree-like random graphs [ISAAC’12, arXiv:1201.1717]: There are empirical evidence that many real-world networks such as Internet or social networks are hyperbolic, and theoretical studies show that hyperbolic networks allow efficient decentralized routing behavior. We study the hyperbolicity of Kleinberg small-world random graphs and a family of tree-like random graphs. Our results show that Kleinberg small-world graphs are not strongly hyperbolic, which indicates that efficient decentralized routing does not imply graph hyperbolicity.
Networked game theory and economics
– Sybil-proof mechanisms in query incentive networks [EC’13, arXiv:1304.7432]: Query incentive networks investigate how to design incentives to allow queries to be propagated to a large fraction of networks so as to find the answers to the queries. We study sybil-proof mechanisms for query incentive networks, and show that a class of direct referral mechanisms is both cost effective and avoid users creating fake identities (sybils) in the network.
– Pricing and revenue maximization with social influence consideration [WINE’11, arXiv:1007.1501]: We study the revenue maximization problem in selling a digital product in a social network where social influence affects agents’ purchasing decisions.
– Game-theoretic approach to overlapping community detection [DMKD’10 special issue on ECML PKDD’10]: We propose a community formation game and use its equilibrium to detect overlapping communities in social networks.
– Bounded budget betweenness centrality game [ESA’09, MSR-TR-2008-167, MSR-TR-2009-78]: We study a strategic network formation game in which nodes strategically select connections to other nodes to maximize their betweenness centralities in the network.