Teaching: Data Science
Research interests: Data-Driven Decision Making, Large-Scale Machine Learning, Web Mining, Social Network Analysis.
Office: Jacobs building, room 505
Ron Bekkerman joined the Department of Information and Knowledge Management as a Senior Lecturer (Assistant Professor) in October 2013. In 2012-2013 he served as Chief Data Officer of Carmel Ventures, the leading Israeli VC fund. He is an Advisory Board member of a number of Israeli startup companies. Prior to that, Ron worked as a Senior Research Scientist at LinkedIn, where he was among the founding members of LinkedIn’s Data Science team. Before LinkedIn, he was a Research Scientist at HP Labs in Palo Alto, CA.
Over the past 14 years, Ron’s research spanned the areas of Data Mining and Machine Learning, which aim to create novel models and algorithms suitable for a variety of practical tasks in statistical data analysis. Currently, he is refocusing his research towards the emerging area of Data Science, where the emphasis is on solving specific business-related problems by applying and adapting Big Data analytics techniques.
Ron’s earlier research work was on text categorization. One of his text classification projects was an initial study of email categorization on the Enron Email Dataset. In 2004 he was among the first researchers to work on mining online social networks. Since then, Ron investigated various setups of data clustering, including multimodal clustering, semi-supervised clustering, interactive clustering, consensus clustering, and one-class clustering. Lately, Ron concentrated his research on scalability of Data Mining in the Big Data era.
While working at LinkedIn, Ron was involved in connecting data analytics and business decision making. He worked on analyzing product virality, assessing user’s lifetime value, and identifying most valuable users. Ron’s data analytics projects contributed to a substantial increase of LinkedIn’s revenues. At Carmel Ventures, Ron got further exposed to a variety of business aspects of technological innovation, such as market size evaluation, go-to-market strategy, and pricing.
Ron holds a PhD in Computer Science from the University of Massachusetts at Amherst, and MSc/BSc degrees (both in CS) from the Technion – Israel Institute of Technology. He is a coauthor of 22 research publications and 5 patent applications. Most of Ron’s work was published in top-tier venues such as JMLR, SIGKDD, WWW, ICML, ECML, SIGIR, CVPR, IJCAI, CIKM, and EMNLP. He has served on Program Committees of 19 top-tier Computer Science conferences and reviewed for 10 international journals. In 2012 and 2013 he served on the Organizing Committee of SIGKDD. Ron is a corresponding coeditor of the book “Scaling Up Machine Learning” published by the Cambridge University Press in 2012.
A Data Scientist's Guide to Making Money from Start-ups (KDD 2013 Panel)
Scaling up Machine Learning (LinkedIn Tech Talk 2012)
Machine Learning: the Basics (LinkedIn Tech Talk 2012)
Improving Clustering Stability (KDD 2009)
S. Budalakoti and R. Bekkerman. Bimodal Invitation-Navigation Fair Bets Model for Authority Identification in a Social Network. In Proceedings of WWW 2012
R. Bekkerman, and M. Gavish. High-Precision Phrase-Based Document Classification on a Modern Scale. In Proceedings of KDD 2011 ppt
R. Bekkerman, M. Scholz, and K. Viswanathan. Improving Clustering Stability with Combinatorial MRFs. In Proceedings of KDD 2009
R. Bekkerman and M. Scholz. Data Weaving: Scaling Up the State-Of-The-Art in Data Clustering. In Proceedings of CIKM 2008
R. Bekkerman. Combinatorial Markov Random Fields and their Applications to Information Organization. PhD Thesis 2008, University of Massachusetts at Amherst ppt
R. Bekkerman, H. Raghavan, J. Allan, and K. Eguchi. Interactive Clustering of Text Collections According to a User-Specified Criterion. In Proceedings of IJCAI 2007 ppt
R. Bekkerman, S. Zilberstein, and J. Allan. Web Page Clustering Using Heuristic Search in the Web Graph. In Proceedings of IJCAI 2007 ppt
R. Bekkerman, K. Eguchi, and J. Allan. Unsupervised Non-topical Classification of Documents. CIIR Technical Report IR-472 2006
R. Bekkerman and M.Sahami. Semi-supervised Clustering using Combinatorial MRFs. In Proceedings of ICML 2006 Workshop on Learning in Structured Output Spaces ppt
R. Bekkerman, R. El-Yaniv, and A. McCallum. Multi-way Distributional Clustering via Pairwise Interactions. In Proceedings of ICML 2005 ppt
R. Bekkerman and A. McCallum. Disambiguating Web Appearances of People in a Social Network. In Proceedings of WWW 2005 ppt
R. Bekkerman, A. McCallum, and G. Huang. Automatic Categorization of Email into Folders: Benchmark Experiments on Enron and SRI Corpora. CIIR Technical Report IR-418 2004
A. Culotta, R. Bekkerman, and A. McCallum. Extracting Social Networks and Contact Information from Email and the Web. In Proceedings of CEAS 2004
R. Bekkerman. Word Distributional Clustering for Text Categorization. MSc Thesis 2003, Technion - Israel Institute of Technology ppt
R. Bekkerman, R. El-Yaniv, N. Tishby, and Y. Winter. Distributional Word Clusters vs. Words for Text Categorization. In Special Issue on Variable and Feature Selection of JMLR 2003
R. Bekkerman, R. El-Yaniv, N. Tishby, and Y. Winter. On Feature Distributional Clustering for Text Categorization. In Proceedings of SIGIR 2001 ppt
Program Committee Member: SIGIR-14, ICML-14, WSDM-14, SIGIR-13, ICML-13, WSDM-13, SIGIR-12, ICML-12, AAAI-11, ACL-11, KDD-10, ICML-10, KDD-09, ECML-09, NAACL-HLT-09, UMAP-09, SIGIR-08, ECML-08, ICML-07.