BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:Dynamic Pattern Recognition in Large-scale Graphs with Application
 s to Social Networks
DTSTART:20201221T170000
DTEND:20201221T190000
DTSTAMP:20260407T014453Z
UID:c7e75b859f3b722a885a56f8df2e7c07b8c870862dc35e4f05824b0c
CATEGORIES:Miscellaneous
DESCRIPTION:Volodymyr Miz\nPublic defense.\n\nA graph is a versatile data 
 structure facilitating representation of interactions among objects in var
 ious complex systems. Very often these objects have attributes whose measu
 rements change over time\, reflecting the dynamics of the system. This gen
 eral data framework can be used in many fields to represent complex data s
 tructures: brain networks and neuronal spikes\, web networks and clickstre
 ams\, social networks and activity of the users\, among others. In all of 
 these examples\, the structural and dynamic components of the data are ins
 eparable\, which significantly complicates the detection\, analysis\, and 
 interpretation of patterns that emerge in the networks. The increasing siz
 e and complexity of graph-structured data require scalable and interpretab
 le algorithms for dynamic pattern detection in such systems.\n\nIn this di
 ssertation\, we present an unsupervised approach for dynamic pattern detec
 tion in large-scale graphs. In this approach\, we combine intuitions deriv
 ed from attention mechanisms\, Hopfield networks\, and memory networks to 
 build scalable\, efficient\, and interpretable algorithms. We then demonst
 rate multiple applications of this approach in recommendation systems\, in
 formation recovery algorithms\, and collective behavior studies. Additiona
 lly\, we use our algorithm to detect dynamic activity patterns in social a
 nd communication networks. We conduct extensive experiments on Wikipedia d
 ata\, detecting and analyzing patterns in the viewership activity in its w
 eb network. To study the collective behavior of Wikipedia readers\, we dev
 elop an automated pattern interpretation model\, which allows for comparis
 on of trending topics across multiple language editions of Wikipedia. The 
 results of the experiments reveal provocative insights into how people int
 eract and search for information in online social networking environments\
 , opening new avenues for future research on collective behavior analysis 
 at a large scale.\n\nFinally\, we present a distributed data processing fr
 amework for Wikipedia server logs that allows others to reproduce all patt
 ern detection experiments presented in this thesis and to conduct similar 
 collective behavior studies on the latest data.
LOCATION:Videoconference https://epfl.zoom.us/j/88100460304
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR