PhD Thesis Defense: Pre-Match Personalized Recommendation to Influence Match Outcomes (and Player Engagement) in Player-versus-Player Video Games
Title: Pre-Match Personalized Recommendation to Influence Match Outcomes (and Player Engagement) in Player-versus-Player Video Games
Speaker: Zhengxing Chen, PhD Candidate, College of Computer and Information Science at Northeastern University
Location: Northeastern University, 440 Huntington Avenue, West Village H, 3rd Floor, Room #366, Boston, Massachusetts 02115
In Player-versus-Player (PvP) video games, match outcomes are believed to be closely related to player engagement and revenues made by game companies. Based on individual players' information and game contexts, personalized recommendation is one family of proposed methods, wherein dynamic game contents are recommended to players in order to influence their win rates.
In this thesis, Zhengxing will focus on personalized recommendation techniques in the pre-match stage, the time when game elements determined before a match starts could already have an impact on the final match outcome. Specifically, he will focus on the recommendations of initial items and opponents to influence match outcomes. First, it is a challenging task in certain PvP games to choose winning-effective initial items because there can be a large amount of choices depending on the strategies of the player and the opponent. Zhengxing will propose a personalized recommender to efficiently seek the optimal initial items to maximize the player's win rate. Second, he will propose an opponent recommendation system in which expected match outcomes within recommended opponents optimize all involved players' engagement.
A distinctive feature of personalized recommendation in PvP games is that influencing one's winning probability will adversely change that of the opponent(s) who are human players too. The initial item recommendation system exemplifies how match outcomes can be affected positively from one side, while the opponent recommendation system exemplifies not only how but also how much match outcomes should be influenced when we consider a population of players.
About the Speaker
Zhengxing Chen is a 5th year PhD student working with Professor Magy Seif El-Nasr. His research focus is on personalized recommendation systems for better player engagement in video games. He has broad interests in churn analysis, behavioral clustering, skill modeling, matchmaking and virtual team composition. His research approaches largely revolve around machine learning, data mining and artificial intelligence techniques.
During his PhD study, he has done internships at Electronic Arts, eBay, Google, and Facebook where he has used AI techniques to improve real-world user engagement problems. Prior to coming to Northeastern University, Zhengxing earned his Bachelor of Science degree at the Beijing University of Posts and Telecommunications in China in 2013.
Professor Magy Seif El-Nasr, Associate Professor, Interdisciplinary with the College of Arts, Media, and Design (CAMD) and the College of Computer and Information Science (CCIS) at Northeastern University (Advisor)
Professor Seth Cooper, Assistant Professor, College of Computer and Information Science (CCIS) at Northeastern University
Professor Christopher Amato, Assistant Professor, College of Computer and Information Science (CCIS) at Northeastern University
Professor Yizhou Sun, Assistant Professor, Department of Computer Science at the University of California at Los Angeles (UCLA)
Tuesday, August 29, 2017 at 11:00am