You are invited to meet a researcher focuses on recommender systems, social network analysis, personalization, user modeling, and machine learning, at the student research conference IIT.SRC 2021
April 22, 2021 at 8:15 am
online via webex: bit.ly/3aoyHub
Event number: 121 072 9272
Peter Dolog: Understanding and Predicting User Behavior by Data Analysis and Machine Learning
(Aalborg University, Denmark)
Almost any human activity has been affected by digitization. We interact with each other by means of e-mail or social networks, we seek and publish information online, we share our opinions online, we shop online, and we pay by credit cards or by mobile payment solutions, and so on. Digitization simplified our lives and made parts of our life more effective. By such interactions online, we are also leaving our trails behind. These trails in the logs and databases are signals of our preferences, interests, needs, and not least partial signals of our behavior. Current digitized world with traces of our activities online provide rich data and gives us possibilities and opportunities to help people at different levels of activities. Computational user modeling gives us possibilities to understand the user behavior and preferences from online electronic data. Personalization models and algorithms help users in their tasks and predict their behavior based on learned user model. In this talk, I will touch upon some approaches we have developed. First, I will start with recent work on context aware recommendation where we developed multi-view latent factor models based on interaction data with a recommender system. I will present our work on joint collective matrix factorization and collective neural embedding as computational/machine learning models for predicting user preference in user context. I will further touch upon impact of ratings on side information on recommendation system algorithms. I will also present the work where we studied impact of user activities in personalized word clouds as means for navigation in social media. I will also try to sketch some interesting directions for future work.
Peter Dolog is an Associate Professor at the Department of Computer Science, Aalborg University, where he currently coordinates the Database and Web Technologies group. He is a faculty member there since September 2006. Previously, he worked at L3S Research Center at the University of Hannover together with Wolfgang Nejdl between 2002 and 2006, where he obtained a Dr. rer. nat. (PhD) degree in computing in March 2006 with summa cum laude on the topic Engineering Adaptive Web-Based Systems. He held visiting positions of different duration at Yahoo Labs Barcelona, Spain (2015, visiting professor), VU Amsterdam, the Netherlands (2005, visiting researcher), Politecnico di Milano, Italy (2004, visiting researcher), and DFKI Kaiserslautern (2003, visiting researcher). He also held a position at the Slovak University Technology (2000–2002). Prior to that, he obtained there his BSc. (1998) and MSc. (2000) degrees in computing (software engineering). His main research interests are recommender systems, social network analysis, personalization, user modeling, and machine learning. In the past, he also contributed to web engineering. He is senior member of the ACM and member of IEEE Computer Society. He was a general (co-)chair of ICWE 2013, UMAP 2014, and ACM HT 2017 conferences.