BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//132.216.98.100//NONSGML kigkonsult.se iCalcreator 2.20.4//
BEGIN:VEVENT
UID:20250802T102256EDT-6042a9nJpr@132.216.98.100
DTSTAMP:20250802T142256Z
DESCRIPTION:Applied Mathematics Seminar\n\nTitle: Multi-view Unsupervised a
 nd Semi-Supervised Clustering based on Content and Connection Information.
 \n\nAbstract: Constrained Low Rank Approximation (CLRA) is a powerful fram
 ework for a variety of important tasks in large scale data analytics such 
 as topic discovery in text data and community detection in social network 
 data. In this talk\, a hybrid method called Joint Nonnegative Matrix Facto
 rization (JointNMF) is introduced for latent information discovery from mu
 lti-view data sets that contain both text content and connection structure
  information. The method jointly optimizes an integrated objective functio
 n\, which is a combination of the Nonnegative Matrix Factorization (NMF) o
 bjective function for handling text content/attribute information and the 
 Symmetric NMF (SymNMF) objective function for handling relation/connection
  information. An effective algorithm for the joint NMF objective function 
 is proposed utilizing the block coordinate descent (BCD) method.\n	The prop
 osed hybrid method simultaneously discovers content associations and relat
 ed latent connections without any need for post-processing or additional c
 lustering. In addition\, known partial label information can be incorporat
 ed into a JointNMF for semi-supervised clustering framework. The experimen
 tal results from several real-life application problems illustrate the adv
 antages of the proposed approaches.\n\nTo register contact : appliedsemina
 rs [at] math.mcgill.ca Join Zoom Meeting https://us06web.zoom.us/j/8532731
 0903?pwd=SlhEak53S2xrNkVYKzl4YUd5KzBudz09 Meeting ID: 853 2731 0903 / Pass
 code: 383854\n
DTSTART:20211206T210000Z
DTEND:20211206T220000Z
SUMMARY:Haesun Park (Georgia Institute of Technology)
URL:/mathstat/channels/event/haesun-park-georgia-insti
 tute-technology-335383
END:VEVENT
END:VCALENDAR