Today

1. CmpE 579 Seminar: Finite automata over matrix groups by Özlem Salehi Köken
   • Start time: 12:00pm, Tuesday, April 24th
   • End time: 01:00pm, Tuesday, April 24th
   • Where: AVS Conference Room, BM

   • Abstract: The group automaton (finite automaton over a group) is a nondeterministic finite automaton model equipped with a register which holds an element from a group. The register is initialized to the identity element of the group, and an input string is accepted if after completely reading the string the machine enters an accept state and the register is equal to the identity element after being multiplied by a group element at every step. In this talk, I will introduce group automata defined by matrix groups and discuss some of our recent results on the subject. We will focus on various matrix groups and demonstrate a hierarchy between the corresponding language classes. 

   • View this event in Google Calendar

Wednesday, May 2nd

1. CmpE Seminar: Predicting Unspoken Views by Kareem Darwish, PhD
   • Start time: 02:00pm, Wednesday, May 2nd
   • End time: 03:00pm, Wednesday, May 2nd
   • Where: AVS Conference Room, BM

   • Abstract: This work examines the effectof online social network interactions on future attitudes. Specifically, wefocus on how a person's online content and network dynamics can be used topredict future attitudes and stances in the aftermath of a major event. In thisstudy, we focus on the attitudes of US Twitter users towards Islam and Muslimssubsequent to the tragic Paris terrorist attacks that occurred on November 13,2015. We quantitatively analyze 44K users' network interactions and historicaltweets to predict their attitudes.  We provide a description of thequantitative results based on the content (hashtags) and network interaction(retweets, replies, and mentions). We analyze two types of data: (1) we usepost-event tweets to learn users' stated stances towards Muslims based onsampling methods and crowd-sourced annotations; and (2) we employ pre-eventinteractions on Twitter to build a classifier to predict post-event stances. Wefound that pre-event network interactions can predict someone’s attitudestowards Muslims with 82% macro F-measure, even in the absence of prior mentionsof Islam, Muslims, or related terms.  We also interrogated our classifierto ascertain the most discriminating features.  Doing so allowed us: todiscover underlying groups that may express similar attitudes on a topic, butmay differ ideologically; and to associate political views with lifestylechoices.                                                  Short Bio: Kareem Darwish is a senior scientist at the Arabic LanguageTechnologies at QCRI with interest in information retrieval, natural languageprocessing, and social computing. He was as a researcher at the Cairo MicrosoftInnovation Lab and the IBM Human Language Technologies group in Cairo. He alsotaught in the Electrical Engineering Department at the German University inCairo and the Faculty of Computer and Informatics at Cairo University.

   • View this event in Google Calendar