Dr. Eleni Koutsomitopoulou obtained her PhD in computational linguistics at Georgetown University in Washington DC. She works at the interface of cognitive neurosciences, linguistics, and computer science, where her research focuses on the neural bases of language acquisition, learning and memory, as well as the exploration of neuro-cognitive and computational modelling processes of both unimpaired and impaired natural language. Her past affiliations include the Editorial and Indexing Group at Reed Elsevier Lexis Nexis USA and Europe, and the Research and Development Group at Reed Elsevier Lexis Nexis USA. She is the co-inventor of Fact Extractor, a U.S. patented took-kit for shallow and deep parsing and lexical, semantic and syntactic annotation of XML input aiming at meaning discovery, as well as entity and fact extraction in unstructured data. She has worked in the development of pattern-matching grammars for English, French, Spanish and German for information extraction, data mining and indexing applications. Eleni has been affiliated with the Summer Institute of Linguistics at UND as well as the Computational Linguistics Group at Trinity College Dublin. Her postdoctoral research focuses on neurocognitive modeling with applications to the following two distinct areas: modeling the neurocognitive processes of speech retraining for adults with acquired hearing loss and modeling the dynamics of belief update and belief revision for purposes of knowledge representation and reasoning.
Research Interests
Computational cognitive modeling of typical and atypical speech and language processes, as found in clinical populations such as hearing impairment, and other communication disorders.
Symbolic and statistical methods for language learning and production.
Information Extraction and knowledge acquisition from large linguistic data sets.
Perception and multi-modal aspects of speech and language processing; Signed Languages.
Machine Translation and Corpus Linguistics methods for MT.
Interaction of procedural and declarative (sub)systems for language processing.
Issues on the integration of unknown lexical and grammatical linguistic information; acquisition of L2
Knowledge representation and belief revision: how the human cognitive system deals with contradictory facts
Exploring the correspondence of fMRI and (CL)AR(NE)T patterns of activation
Current work (in prep.):
Grammar-driven pause prediction in clear and casual speech.
Modeling the acoustic learning problem on an ART-type neural network: speech retraining for adults with hearing loss
Turn-taking affected by differentiated rhythmic patterns in speech of adults with hearing loss.
King-Kopetzky Syndrome or 'obscure auditory dysfunction': modeling auditory dysfunction in the absence of sensory (auditory) disability
Past research:
Development of computational tools and methods for the recording, analysis, annotation, research, machine translation solutions and teaching of Greek sign language (GSL)
A CLARNET model of belief revision and belief update [in collaboration with Dr. Carl Vogel, Trinity College Dublin Computational Linguistics Group]
Integration of UNKs using "reinforced perception techniques" (TM)
Is the declarative/procedural difference categorical? A view from within an adaptive resonant semantic network
Using "reinforced perception techniques" for the mitigation of lateness in sensory input associated with cases of inferior sensory acuity (aka sensory disabilities
Skills
Programming: C/C++, LISP, Prolog, Perl, Python, UML, HTML, LaTeX, XML. Mac OS X, Windows, Linux
Productivity software, as well as web servers, and tools and libraries related to computational linguistics.
Languages: Greek (native), English (fluent), French (intermediate), German (basic), Spanish (basic)
Signed Languages: ASL, ISL, GSL
Education
Doctor of Philosophy in Computational Linguistics May 2004 [PhD awarded with distinction]
Georgetown University Washington, DC [Research focus: Neurocognitive computational modeling for the representation of natural language, Citation: A neural network model for the representation of natural languageby Koutsomitopoulou, Eleni Ph.D., Georgetown University, 2004, 522 pages; UMI order number/AAT 3137058] Departmental Assistantships Fulbright Scholarship
Master of Arts in Linguistics February 1996
University of Crete Crete, Greece [Focus areas: Greek Language and Linguistics] European ERASMUS Scholarship for Research and Teaching Assistantship
Bachelor of Arts February 1992
University of Athens Athens, Greece [Focus areas: Philosophy Education Psycholinguistics]
Honors Fulbright Doctoral Scholarship [July 1997] ERASMUS European Scholarship University of Amsterdam (Netherlands) [Sept 1995-Feb 1996] ERASMUS Research Scholarship University of Crete (Greece) [Sept 1993-Aug 1995]
Additional Training
Critical Discourse Analysis Studies. Sept 1995 - Feb 1996. Universiteit van Amsterdam (The Netherlands)
Corpora and Translation Course. Dec 13-15, 2004. The Center for Corpus Research. University of Birmingham (Birmingham, England)
Major Past Research Projects
Proper-Name Entity Recognition in multilingual News, Business, Legal and Financial data
Fact Discovery and Extraction in multilingual News, Business, Legal and Financial data
Lisp-based semantic interpreter for the detection and extraction of relationships among entities and facts in pharmacological data
Grammar development for in-house Fact-Extraction system applied to a variety of document sources such as Executive Movements and Mergers & Acquisitions
Coreference resolution and semantic disambiguation via adaptive learning methods and cognitive modeling
Adaptive learning techniques for automated partitioning of data and improved online searches
Employment
Toshiba Research Europe (Cambridge, UK) Nov 2014-current Research contractor Developing NLP tools for the NLU component of a Spoken Dialogue System R&D on prosodic phrase breaks (aka silent pauses) using a grammar-based prediction model. Binary and multilevel pause prediction model. Related evaluation metrics.
SIL-UND summer 2010 Faculty at the Summer Institute of UND Taught (LING-507) Computational Syntax and Morphology at the Summer Institute of Linguistics of the University of North Dakota.
TRINITY COLLEGE DUBLIN– The University of Dublin (Dublin, Ireland) March'09 – Aug'10 Research Fellow Research in modeling natural language data for purposes of topic detection, text categorization and belief revision for knowledge representation.
Specialized in Text Analytics and related Natural Language Processing methodologies and techniques.
Developed grammars for entity and fact extraction in a wide range of natural language data (including News, Business, Financial, Legal, Medical/Pharma).
Standardized architecture and design blueprints for the development of custom-tailored grammars.
Provided insight to management and proposed data extraction and indexing business solutions.
Standardized development, training, and testing processes for extraction services in a wide range of products and data.
Responsible for the drafting and dissemination of functional and technical documentation.
Served as primary point of contact for development of entity and fact extraction algorithms and related technology transfer across various groups and business units.
Known for timeliness and efficiency in project completion.
Served as point of contact for the evaluation of new tools and technologies.
Developed, maintained, installed, enhanced and optimized indexing and information extraction processes as well as metadata for products and technologies.
Applied computational linguistic theories, methodologies, and techniques for the execution of algorithms for the extraction of entities and facts from natural language data.
Contributed to the successful development of high-quality products and processes by effectively tracking, analyzing and reporting software design problems and related code bugs.
Proposed solutions and executed productivity improvements, contributing toward business objectives.
Communicated effectively, tactically and strategically with internal and external customers, management and functional areas of the company.
Developed rule-based pattern matching grammars for information extraction in English and French natural language data. Designed and developed regression testing techniques and related algorithms for QA purposes.
Provided data analysis and interpretation insights to management regarding indexing and extraction technologies.
Effectively analyzed natural language data to support current/future indexing products and application planning.
Carried out bug reports and performed related analyses and debugging to successfully support product QA.
AMBERCON., INC. (Fairfax, Virginia) 1998 – 2000 Computational Linguist Conducted natural language processing (NLP) research. Contributed to the development of in-house LISP parser. Quality-assurance testing and development of data and related computational resources.
Patent FactEXctrator, LexisNexis USA, United States Patent 20050108630): a “tool-kit” that includes a shallow and deep-parser, and various annotators for implementing XML input aiming at meaning discovery, entity and fact extraction (role: computational linguist, co-inventor).
Polysemy: Theoretical and Computational Approaches. Oxford University Press. Review for Linguist List: http://cf.linguistlist.org/cfdocs/new-website/LL-WorkingDirs/pubs/reviews/get-review.cfm?SubID=14403
Allan, Keith (2001) Natural Language Semantics. Blackwell Publishers. Review for Linguist List: http://cf.linguistlist.org/cfdocs/new-website/LL-WorkingDirs/pubs/reviews/get-review.cfm?SubID=4068
Serving as Reviewer: Angelo Cangelosi, Guido Bugmann & Roman Borisyuk (Eds.), Modeling Language, Cognition and Action. 2005. NCPW9 post-conference proceedings. World Scientific.
Selected Conference Papers and Presentations
"Natural language processing for the prediction of pausing patterns in clear and casual speech?" (In prep.)
AKRR'05 "A Neurobiological Model of Fact Resolution" International and Interdisciplinary Conference on Adaptive Knowledge Representation and Reasoning at Espoo, Finland 15-17 June 2005.
CSDL 2000. University of California at Santa Barbara, Santa Barbara CA May 12, 2000 “The role of Conceptual Metaphor in Knowledge Engineering: Metaphor-based Ontologies.”
Georgetown University, Washington D.C. Oct 16, 1998 Guest Lecture on American Sign Language and Linguistics.
Book TEACHING MODERN GREEK TO ENGLISH SPEAKERS [May 1997] Modern Greek for Beginners. 1997. F. Kavukopoulos, E. Koutsomitopoulou, M. Argirudi, V. Dagkli. Pub: Nefeli and University of Crete Department of Linguistics Press. Athens, Greece. Καβουκόπουλος, Φ., Κουτσομητοπούλου, Ε., Αργυρούδη, Μ. & Δαγκλή, Β. (1997). Νέα Ελληνικά για Αρχάριους, με ελληνο-αγγλικό λεξικό αρχαρίων. Κρήτη: Νεφέλη, Πανεπιστήμιο Κρήτης - Τμήμα Φιλολογίας. [based on applied linguistic and lexicographical research, in-class teaching and evaluation of original material, includes a Modern Greek Language Dictionary for Beginners]
I support Prof. Stefan Müller's initiative for Open Access to free academic journals and books for people who provide the content, read, buy and use them.
