Omri Abend

PhD in Computational Linguistics
Hebrew University of Jerusalem
Research Topic: Grammatical Annotation Founded on Semantics: A Cognitive Linguistics Approach to Grammatical Corpus Annotation.

Omri is currently a Research Associate (post-doc) in the University of Edinburgh, working with Prof. Mark Steedman. 
 
His field of interest is Natural Language Processing (NLP), the scientific and engineering field that addresses the computational modeling and machine understanding of language. His research tackles the semantic representation of text, a key component in nearly all modern NLP applications. His thesis proposed an approach to semantic representation for language, based on distinctions that are cross-linguistically valid, intuitive for people with no previous background in linguistics and suitable for NLP applications. In his post-doc, He is  using statistical learning methods to learn and refine such semantic representations so to account for the individual characteristics of languages. His intention is to use these representations to improve various NLP applications, focusing on Machine Translation, as well as to address the computational modeling of child language acquisition.
 

A personal perspective:
"The generous support of the Azrieli Foundation, as well as the kind and warm support of the Azrieli staff, allowed me to devote my time and efforts to research. This has ultimately led me to set wider, more ambitious goals to my research. I am truly grateful for having received the opportunity to be an Azrieli Fellow."

 
 

  1. Lexical Inference over Multi-Word Predicates: A Distributional Approach. Omri Abend, Shay B. Cohen and Mark Steedman, ACL 2014 (long paper).
     
  2. UCCA: A Novel Framework for Semantic Representation. Omri Abend and Ari Rappoport, ACL 2013 (long paper).
     
  3. UCCA: A Semantics-based Grammatical Annotation Scheme. Omri Abend and Ari Rappoport, IWCS 2013 (long paper).
     
  4. Learnability-based Syntactic Annotation Design. Roy Schwartz, Omri Abend and Ari Rappoport, COLING 2012 (long paper).
     
  5. Neutralizing Linguistically Problematic Annotations in Unsupervised Dependency Parsing Evaluation. Roy Schwartz, Omri Abend, Roi Reichart and Ari Rappoport, ACL 2011 (long paper).
     
  6. Improved Unsupervised POS Induction through Prototype Discovery. Omri Abend, Roi Reichart and Ari Rappoport, ACL 2010 (long paper).
     
  7. Fully Unsupervised Core-Adjunct Argument Classification. Omri Abend and Ari Rappoport, ACL 2010 (long paper).
     
  8. Type Level Clustering Evaluation: New Measures and a POS Induction Case Study. Roi Reichart, Omri Abend and Ari Rappoport, CoNLL 2010 (long paper). * – Both authors equally contributed to the paper.
     
  9. Unsupervised Argument Identification for Semantic Role Labeling. Omri Abend, Roi Reichart and Ari Rappoport, ACL 2009 (long paper).
     
  10. A Supervised Algorithm for Verb Disambiguation into VerbNet Classes. Omri Abend, Roi Reichart and Ari Rappoport, COLING 2008 (long paper).
"During my years as an Azrieli fellow I led various volunteer programs through the Brera Center, a student volunteering organization that aims to promote public and social leadership. The projects focused on teaching computer skills to children and teenagers from the socio-economically weaker neighborhoods of Jerusalem. Topics taught in the courses include: programming in C++, technical English and “hands-on” science for children. The audience ranged from elementary school children with general interest in science, to high-school teenagers interested in improving their technical skills.
 
I am very grateful for my experiences in the Brera Center. It was very satisfying for me to share the knowledge I have with interested children and teenagers. Perhaps more importantly though, I felt that the courses we organized helped in creating a link between the world of my colleagues and me at the university and the world of these children and teenagers. I hope this experience will motivate those of them interested in scientific and technological studies to pursue this path."