George Popescu

Background

Thesis title

"Information Technology for Negotiable Decisions in Group and Social Recommenders", SNF 138054.

Short biography

George was enrolled as PhD student at EPFL in the HCI Group working under the supervision of Dr. Pearl Pu between 01.10.2010 and 01.10.2012.

His research focused on human factors in computer science, more specificaly:

group and social recommendation technology
computational social choice and welfare
social resources for group recommendations
decision, negotiation, manipulation and collaboration in groups

Before his PhD George was awarded the Masters degree in Complex Informatics Systems from École Polytechnique in France. Starting directly his second year of the program he received an "Excellence scholarship" from the French Government. While studying in Paris he won the 1st prize at the Robot Challenge anual competition against École Normale Supérieure. His degree culminated with his Master thesis defense after working as an intern on "Management of IT applications" at the headquarters of Siemens Wind Power in Denmark for 6 months.

Prior to this, George obtained his Bachelor degree in Computer Science from the Polytechnics University of Bucharest, Romania, with the paper "Voice command Kephera III". In the final report he developed a voice recognition software which helps users control the path of Kephera III robots using their voice wirelessly. The application displayed both the phrases pronounced by users as well as key words extracted to move the robot in a certain way, e.g.: "left", "right", "back", "front", "slower", faster", "turn" and even "happy". During the four years he received a "Performance scholarship" from the university and participated in many students' scientific contests for which many of his papers have been awarded highest distinctions.

Education

10/2010-10/2012 | PhD in Computer Science – Human Computer Interaction Group, École Polytechnique Fédérale de Lausanne, Switzerland
08/2009-09/2010 | MSc in Conception and Management of Complex Informatics Systems - École Polytechnique & Telecom ParisTech, Paris, France
10/2005-07/2009 | BSc in Computer Science, Polytechnics University of Bucharest, Romania

Employment

10/2010-Present | Doctoral assistant, School of Computer and Communication Sciences, EPFL, Switzerland.
03/2010-09/2010 | IT Intern, Siemens Wind Power A/S, Denmark.
09/2008-07/2009 | Technology Consultant, Oracle Corporation, Romania.
06/2008-09/2008 | IT Networking trainee, Orange, Romania.

Research

Interests

Current interests connect multi-disciplinary approaches to social and group recommender systems such as: human computer interaction, behavioral decision theories, social choice and user experience.

Previous work

The success of a group is strongly related to its members’ satisfaction while making common decisions. Using an eye-tracking device we carried on experiments to show that people do not have inate preferences but change according to group dynamics and group influence. The main results prove that the characteristics of the group, namely member familiarity and trust, dictate the overall group satisfaction and decision acceptance. In the more connected group participants not only change their preferences more being more influenced by other members, but they are also more satisfied with the overall outcome. These in-depth results are the first to correlate eye-tracking data and preference dynamics to understand decision making induced by social influence.

During previous work at the Human Computer Interaction laboratory one of the major contributions was the demonstration of the applicability of the probabilistic weighted sum (PWS) algorithm for group recommendation strategies and negotiation. We analyzed different group recommendation approaches with respect to group satisfaction and discussed key satisfaction issues to be taken into account. The PWS algorithm we proposed computes probabilities instead of fixed scores for songs to appear in groups’ playlists favoring music diversity and the discovery of new items.

GroupFun is our web application that helps a group of friends to agree on a common music playlist for a given event they will attend, e.g. a birthday party or a graduation ceremony. Firstly, it is implemented as a Facebook plugin connecting users with their friends. Secondly, it is a music platform that helps individuals to manage their favorite songs in groups. In GroupFun users can listen to their own collection of songs as well as their friends’ music. With the collective music database, the application integrates friends’ music tastes and recommends a common playlist to them. Therefore, the application aims at satisfying music tastes of the whole group by aggregating individual preferences through the use of the PWS algorithm.

Screenshots are available here:
http://hci.epfl.ch/projects/GroupFun/index.php

Keywords: group recommender, human computer interaction, social choice, negotiation, collaboration, algorithms, human factors, satisfaction, group decision, voting.

Other courses

06/2012 EPFL Summer Research Institute
Spring 2012 - PhD assistant for the Human Computer Interaction course
Autumn 2011 - PhD assistant for the Informatics course
06/2011 EPFL Summer School in Computer Science
05/2009 Summer School in Automatics and Computer Science – Polytechnics University of Bucharest
02/2009-07/2009 Adaptive Learning Spaces – UML training for software analyst, National College of Ireland, Fraunhofer Institut Experimentelles Software Engineering, Institut National Polytechnique de Toulouse, Softwin
06/2008-09/2008 CISCO Certified Courses - CCNA1 & CCNA2
10/2006-05/2007 DB and RDBMS – Oracle diploma

Publications and conferences

Papers

Popescu, G. and Pu, P. Eye-tracking Group Influence - Experiment Design and Results. EPFL Technical Report. Lausanne, VD, Switzerland, September, 2012. [pdf]
Popescu, G. and Pu, P. What’s the Best Music You Have? Designing Music Recommendation for Group Enjoyment in GroupFun, ACM SIGCHI Conference on Human Factors in Computing Systems, Works-in-Progress, Austin, Texas, USA, May 2012. [pdf]
Popescu, G. and Pu, P. Probabilistic game theoretic algorithms for group recommender systems, ACM RecSys, Workshop on Music Recommendation and Discovery, Chicago, IL, October 2011. [pdf]
Popescu, G. Preference aggregation in group and social recommender systems, Candidacy exam paper, EPFL, August 2011. [pdf]
Popescu, G. Validation of game theoretic algorithms for recommender systems, Semester report, EPFL, July 2011. [pdf]
Popescu, G. Group recommender systems as a voting problem, Semester report, EPFL, February 2011. [pdf]
Master thesis: "Management of IT applications at Siemens Wind Power A/S" internship report, Paris, 2010.
Bachelor thesis: "Voice command Kephera III", Bucharest, 2009.

Awards

02/2010 1st Prize Robot challenge: annual competition between École Polytechnique and École Normale Supérieure (autonomous robot conception and architecture design), with Perciu A, Pavel A-M, Spasova I, Fang Z.
05/2008 3rd Prize, Polytechnics University of Bucharest, Scientific Students Session - paper: “Contributions regarding knowledge management in multi-agent systems”, with Balan C.
05/2006 Special Prize, Polytechnics University of Bucharest, Scientific Students Session, Mathematics - paper: “Binary Codification of Forms”
05/2007 4th Prize, Polytechnics University of Bucharest, Scientific Students Session - paper: “Event methods for traffic management”
05/2006 1st Prize, Polytechnics University of Bucharest, Scientific Students Session, Mathematics - paper: “Binary Codification of Images”

Events

06/2011 & 06/2012 I&C Research Day @ EPFL: Demo and poster session
09/2011 LUXS 2011: Landscape of User Experience in Switzerland

Personal

Associations

03/2012-Present A/RO Romanian Student Association - President (10/2010 - 03/2012 Committe Member)
08/2011-Present GSA I&C Graduate Student Association - Committee Member
07/2011-Present Mensa The International High IQ Society - Member
10/2010-Present ACIDE EPFL Association of Research Scientists and Lecturers - Committee Member
10/2010-Present ACM Association for Computing Machinery - Reviewer
09/2006-09/2010 AIESEC (Bucharest, Paris, Aarhus) - Information Manager
10/2009-03/2010 AMIX Masters Association of École Polytechnique - Communication Manager

Hobbies

Summer: Tennis, Table tennis, Swimming, Sailing and Windsurfing

Winter: Snowboarding, Skiing

Travelling: Experiencing distant cultures

My favorite sport is lawn tennis which I have been practicing since the age of 12 and am always looking forward for a new match. I also enjoy playing table tennis, football and bowling. After discovering a new passion for snow and water sports I became addicted to snowboarding, skiing, swimming, sailing and windsurfing. Travelling and interacting with people from different cultures allowed me to understand distant cultures and share experiences.

George Popescu, PhD Student