CLUK Database of Researchers

Researchers' names are underlined with either their webpage url or email.
Please click for further information or contact.

Members:

John Barnden    <J.A.Barnden@cs.bham.ac.uk> 
Univ. of Birmingham: School of Computer Science, Birmingham, West Midlands, B15 2TT, U.K.,
Research keywords: pragmatics   user modelling   text analysis  

I am interested in systems for reasoning about agents' beliefs and reasoning, and in systems that do metaphor-based reasoning. In particular, I'm interested in metaphors for mental states and processes. I have implemented a system for metaphor-based (and other) reasoning about beliefs and reasoning.  

Current projects: (i) Controlling ambiguity and style in NLG, (ii) Learning probabilistic context-free grammars.
General: theory of automata, formal language learnability, natural language learning, genetic algorithms and computational search in general, finite-state methods for NLP, statistical language modelling

I am interested in linguistic approaches to machine translation, and also in formal approaches to linguistic description (especially LFG and HPSG).

I'm interested in developing a computational model of quantitative phonological research (from the field to the laboratory).  

Kalina Bontcheva    <kalina@dcs.shef.ac.uk> 
University of Sheffield, Regent Court,211 Portobello St, Sheffield, S. Yorkshire, S1 4DP, U.K.,
Research keywords: generation   user modelling   pragmatics  

Nadjet Bouayad-Agha    < nadjet@itri.brighton.ac.uk> 
ITRI, Watts Building, Moulsecoomb, Brighton BN24GJ
Research keywords: generation   multimodal systems   text analysis  

I am working at improving the expressive power of generation systems by making it possible to vary the layout of the output texts. I am also interested in theories of discourse, style and in using corpora.

My research concerns extraction of knowledge (facts, information) from tagged text. My program, KEP, attempts (amongst other things) to build glossaries totally automatically. This involves extraction of acronyms/abbreviations, technical terms, and (most difficult of all) definitions/examples/hypernyms etc. I also specialise in the development of standalone functions to perform specific tasks, such as returning the singular form of a plural English noun (without recourse to MRDs, so that the 'C' function may be embedded in any program that requires it).  

This has focused on the area of Intelligent Tutoring Systems (ITS) for Language Teaching or Intelligent Comuter-Aided Language Learning (ICALL). The bulk of research has been concentrated on building and extending LICE (Language Independent Composition Environment) a prototype ICALL system for teaching writing to University level students of German. Related areas of interest therefore include natural Language Processing (handling ill-formed input (at the levels of text, syntax and semantics), processing tutorial dialogue and evaluating student help requests), Parsing, Generating, Expert and Knowledge-based systems for capturing knowledge about teaching and learning as well as student modelling. Recent research has been in the areas of producing interactive tutorial dialog, handling ill-formed input by means of genetic algorithms and building ITS for the WWW.  

Andrew Bredenkamp    <andrewb@essex.ac.uk> 
University of Essex, Wivenhoe Park, Colchester, Essex, CO4 3SQ, U.K.,
Research keywords: machine translation   engineering systems   syntax  

I am interested in all aspects of the industrial use of language engineering technologies, particularly multilingual langugage engineering and machine translation.  

Ted Briscoe    <Ted.Briscoe@cl.cam.ac.uk> 
University of Cambridge
Computer Laboratory, William Gates Building, JJ Thomson Avenue Cambridge CB3 0FD, England
Research keywords: lexicons   parsing   grammar induction  

see web page  

Lynne Cahill    <lynneca@cogs.susx.ac.uk> 
University of Sussex, Falmer, Brighton, BN1 9QH, U.K.,
Research keywords: lexicons  

Gerald Gazdar (Sussex) and I are working on a trilingual lexicon of English, German and Dutch, aiming to exploit the similarities between them. See Gerald Gazdar's entry for more details.  

Jo Calder    <J.Calder@ed.ac.uk> 
University of Edinburgh, 2 Buccleuch Place, Edinburgh, EH8 9LW , U.K.,
Research keywords: parsing   semantics   teaching  

My research consists of several interconnecting themes: what methods are appropriate for the formulation of descriptions of language? I have looked at various kinds of feature-value formalisms and their use in encoding theories of grammar such as categorial grammar. I have investigated extensions of such formalisms and theories to areas such as morphology, phonology and the lexicon. how can descriptions of language be exploited computationally? I have worked on the ESPRIT projects Acord, on natural language and graphical interfaces to databases, Dyana (DYANA-2 on the dynamic interpretation of natural language and Dandelion Dandelion on linguistic and computational theories of discourse. In the course of this work, I have produced a `grammar development shell' which hosts a variety of grammatical formalisms. what computational support can be given to grammar writers? I have looked at ways of visualizing derivations for the purposes of grammar debugging and the teaching of parsing techniques. I have also designed and developed a parameterizable Diagram Editor called Thistle.  

Paul Callaghan    <p.c.callaghan@durham.ac.uk> 
University of Durham, Durham, U.K.,
Research keywords: information extraction   machine translation  

Formerly working with the Durham LNLE LOLITA project: my thesis is on evaluation, in particular the recent MUC-6 and LOLITA's performance in it. Currently working in Durham on the "Mathematical Vernacular" project - we are studying so-called "mathematical language" with a view to implementing an interface to type-theoretic proof checking. See my web page for more details!  

Yvonne Canning    <Yvonnecanning@aol.com> 
University of Sunderland, School of Computing & Information Systems, St Peter's Wat, Sunderland, Tyne & Wear, SR6 0DD, U.K.,
Research keywords: syntax   generation   text analysis  

I received my PhD in April 2002 for a thesis entitled "Syntactic Simplification of Text". I am interested in: syntax, anaphora resolution and replacement, summarization, simplification, activisation of passive clauses, splittting of compound sentences.  

Main interests: large-scale grammar and lexicon development, practical NL parsing. Theoretical frameworks: unification-based phrase-structure grammars, lexicalised tree-adjoining grammar formalisms. Applications of interest: extraction of lexical information from corpora, robust parsing for multilingual information retrieval, text simplification for people with comprehension deficits (aphasics).  

Alexander Clark    <alexc@cs.rhul.ac.uk> 
Department of Computer Science, Royal Holloway University of London, Egham, Surrey TW20 0EX United Kingdom
Research keywords: parsing   grammar induction  

Grammatical inference, statistical machine learning, first language acquisition, morphology. address:

Statistical NLP; statistical parsing; using WordNet for estimating probabilities of lexical dependencies; categorial grammar  

COBUILD maintains a corpus of contemporary English from a wide range of spoken and written sources and analyses this corpus through computational and manual methods to prepare dictionaries and other language reference products.  

John Coleman    <john.coleman@phon.ox.ac.uk> 
Oxford University Phonetics Lab, 41 Wellington Square, Oxford, OX1 2JF, U.K.,

computational phonology; experimental phonetics; neurolinguistics  

My research interests cover the broad area of empirical computational linguistics and are currently focussed on the application of (supervised) machine learning for helping users access (text-based) information. Information access covers a broad spectrum of technologies, which in my case means information extraction, retrieval and machine translation. Previously I've looked at IE applied specifically to molecular biology and from now I will be looking at domain-portable IE systems as well as related issues such as training set design, need for user dialogue, integration of robust rule-based and ML-based technologies, etc. for this task. I've also worked on cross language IR (English-Japanese) and I maintain a strong interest in working with Japanese and other Asian languages on IR and IE-related tasks.  

Computational semantics and reversible, robust reference resolution. Currently investigating use of a fragment of linear logic as a glue language for providing a formal semantics for semantic underspecification and context update.  

Angelo Dalli    <angelo@maltalinks.com> 
University of Sheffield
Regent Court 211 Portobello Street Sheffield, S1 4DP
Research keywords: text analsyis  dialogue systems   information extraction 

My research has been on the application of bioinformatics algorithms, such as alignment, together with sophisticated adaptive clustering algorithms, to the task of automated lexicon building. I have also created a web-services based framework for the interchange of linguistic data.

George Demetriou's general research interests are in Artificial Intelligence and Language Engineering. More specifically, George's research is related to: methods of Knowledge Acquisition and Representation for Text-Based Intelligent Systems with reference to large-scale applications in the areas of Speech Recognition, Natural Language Understanding and Information Retrieval; statistical and knowledge-based Natural Language Processing and the development of lexical databases from unstructured text resources; tools for natural language analysis and automatic text processing. George's past work includes the investigation of Constraint Satisfaction techniques for word sense disambiguation. Current research focuses on methods of semantic information engineering for large-vocabulary man-machine speech communication systems.  

My interests lie in the simplification of text for aphasic readers. To date, for my PhD research, I have concentrated on lexical simplification by means of synonym substitution, using existing lexical resources such as WordNet and the Oxford Psycholinguistic Database. Currently I am working on an EPSRC project called PSET (Practical Simplification of English Text) which involves researchers from both Sunderland and Sussex Universities, the aim of which is to produce and test a complete system that simplifies on both a lexical and syntactic basis.  

Alexiei Dingli    <alexiei@dcs.shef.ac.uk> 
University of Sheffield
Dept of Computer Science, Regent Court, 211 Portobello Street, Sheffield, S1 4DP. UNITED KINGDOM
Research keywords: information extraction   semantics   synta  

My research is mainly concerned with developing new Adaptive Inforamtion Extraction Algorithms. Information Extraction (IE) is the problem of converting text such as Web pages into structures data objects siutable for automatic processing. So far, IE has been only available in research labs and this due to the complexity and expertise required to train existing IE systems. By Adaptive IE (AIE) we are planning to creat systems adaptable to the user's needs without any need of expertise and whereby the handholding required by such system from the users is requced to an absolute minimum.

Philip Edmonds    <phil@sharp.co.uk> 
Sharp Laboratories of Europe Ltd, Edmund Halley Road, Oxford, OX4 4GB, U.K.,
Research keywords: lexicons   semantics   machine translation  

Research into the representation and use of near-synonyms and the fine-grained differences between near-synonyms (i.e., nuances). I have developed a knowledge-based model in which near-synonyms are differentiated in terms of denotation, style, and expressed attitude within subconceptual clusters (i.e., clusters that are outside of an ontology, yet link to it). I have also developed a statistical model of which of a set of near-synonyms is chosen in context.  

My main area of interest is in the field of unsupervised natural language learning. In particular, the search for generic human and inter-species language universals to devise computational methods by which language can be descriminated from non-language and core structural syntactic elements of unknown languages can be detected. Aims of this research include: contributing to the understanding of language structure and the detection of intelligent language-like features in signals, to aid the search for extra-terrestrial intelligence.  

This work will improve the performance of Internet document search and retrieval by applying a shallow text analysis based on the "Lexical Chains" notion introduced by Morris and Hirst 1991. This will produce a profile that indicates web page contents as strengths in the categories of Roget's thesaurus. This will allow retrieved pages to be clustered according to their similarity, and in relation to a page that has been indicated as exemplifying a particular topic.  

Computational linguistics and language engineering: particularly lexical description (using the lexical description language DATR), multilingual lexicons, information extraction, text parsing and generation, and software tools  

My research covers corpus linguistics, computational linguistics, and natural language processing, including the construction and analysis of linguistic corpora, automatic wordclass tagging, grammar induction, robust syntactic parsing with case-based reasoning, and lexicon construction. I have designed and implemented AUTASYS (http://www.phon.ucl.ac.uk/home/alex/project/tagginging/tagginging.htm), an automatic tagging system, and the Survey Parser (http://www.phon.ucl.ac.uk/home/alex/project/parsing/parsing.htm), both of which have been applied in research into text-to-speech synthesis and speech recognition. The Survey Parser was instrumental in the completion of the British component of the International Corpus of English.  

Frederik Fouvry    <fouvry@essex.ac.uk> 
University of Essex, Wivenhoe Park, Colchester, Essex, CO4 3SQ, U.K.,
Research keywords: robust parsing   engineering systems   machine translation  

See my web page for an overview of the projects I was involved in. My main interest is in usable natural language processing applications and development of these applications.  

Robert Gaizauskas    <r.gaizauskas@dcs.shef.ac.uk> 
University of Sheffield, 211 Portobello Street, Sheffield, South Yorkshire, S1 4DP, U.K.,
Research keywords: information extraction   robust parsing   semantics  

My research interests focus on information extraction (IE) as an applied NLP technology, but range over all aspects of computational linguistics (CL) that have bearing on IE (which includes virtually all of CL). In particular I am interested in discourse representation, coreference resolution, robust parsing and knowledge representation for NLP.  

Gerald Gazdar    <geraldg@cogs.susx.ac.uk> 
Cognitive & Computing Sciences, University of Sussex, Brighton, East Sussex, BN1 9QH, U.K.,
Research keywords: lexicons  

Lynne Cahill (Sussex) and I are developing a trilingual computer lexicon for the core vocabulary of Dutch, English and German. From a linguistic perspective, we are ascertaining the extent to which these Germanic languages can be lexically related, examining formal ways of expressing linguistic generalizations that hold across two or more languages, and assessing the degree to which the historical links between languages can be exploited in descriptions of the languages as they are now. From a computational perspective, we are evaluating how well existing techniques for representing monolingual lexicons generalize to the multilingual case and investigating the extent to which multilanguage lexical representation techniques can be applied within monolingual lexicons. Roger Evans (Brighton), Bill Keller (Sussex) and I have been responsible for the design of a formal language for lexical knowledge representation. DATR is a declarative language for representing a restricted class of inheritance networks, permitting both multiple and default inheritance. The principal intended area of application is the representation of lexical entries for natural language processing. The goal of the DATR enterprise is the design of a simple language that (a) has the necessary expressive power to encode the lexical entries presupposed by contemporary work in the unification grammar tradition, (b) can express all the evident generalizations about such entries, (c) has an explicit theory of inference, (d) is readily implementable, and (e) has an explicit declarative semantics. Development of the language continues. In work with Roger Evans (Brighton) and David Weir (Sussex), I have ben investigating the use of DATR to encode a Lexicalized Tree Adjoining Grammar (LTAG) lexicon as an inheritance hierarchy with internal lexical rules. Such an encoding eliminates the considerable redundancy otherwise associated with an LTAG lexicon.  

Javad Gholamy    <gholamyj@yahoo.com> 
Tehran University, P.O.Box 14155-6553, , Tehran, Janbe Pole Ghisha, 14170, Iran,
Research keywords: text analysis   information extraction   information retrieval  

The present research study is entitled" Criteria for Intralingual Translation". In this project, we hope to reach at some objective techniques & linguistic criteria for discoursal modification of academic resources writtten and predominantly used by native English-speaking scientific communities. Intralingual Translation as a cover term is used to account for all sorts of modifications at phonological, lexical , syntactic, and discoursal levels. The study contrastively analyses Voice of America's Special English programs as an ideal instance of naturally modified ESP passages with thematically similar topics adopted from native-used media releases. Of course, I have come up with a lot of mind-boggling methodological obstacles particulaly in terms of comparing the textual features of modified and non-modoified corpora. And I hope somebody who has a good understanding of text linguistics and right software facilities at his/her disposal to lend me a hand. I do most of the laborious job manually which I believe can be done via computer. All commenents, suggestions, and criticisms are welcomed wholeheartedly on all dimentions of the present study.  

I'm working on argumentative advisors: how to strategically give advice which is tailored to the addressee's perspectives and values, by using a theory of informal argumentation (New Rhetoric). Application domain: healthy nutrition promotion.  

Matthew Hurst    <hurst@etl.go.jp> 
Centre for Cognitive Science, U.K.,
Research keywords: information extraction   robust parsing   engineering systems  

I am interested in the interaction between document layout and language, with particular interest in the extraction of information from tabular presentations.  

Janet Hitzeman    <hitz@linus.mitre.org> 
University of Edinburgh, HCRC/CSTR, 2 Buccleuch Place, Edinburgh, Scotland, EH8 9LW, U.K.,
Research keywords: generation   pragmatics   text analysis  

I am using information concerning the structure of a text (anaphoric, rhetorical, paragraphing, etc.) to improve the quality of speech synthesis systems, i.e., concept-to-speech synthesis.  

To identify software and hardware suitable for logistics, quantify software and hardware within this area, apply new or tried techniques of mobile commnications within the logistics/courier sector.  

Large-corpus analysis, including properties of translation, typology, and morphological analysis techniques.  

My research interests are within the area of text understanding. In particular, my PhD research study involves the development a new candidate story understanding system, called SOPHIA. The philosophy behind the SOPHIA story understanding system has been to improve and extend previous theories of conceptual analysis of story understanding by using a new model of semantic memory that I have developed.In my future research I would like to investigate some of human's cognitive activities during literary aesthetic text and film appreciation.  

I'm mostly interested in `computational semantics', that is the application of results from formal semantics to NL interpretation and generation. Currently I'm involved with the GNOME project, which aims to develop new algorithms for generating nominal expressions informed by empirical work in corpus analysis and experimental psychology.  

Adam Kilgarriff    <Adam.Kilgarriff@itri.bton.ac.uk> 
ITRI, University of Brighton, Lewes Road, Brighton, Sussex, BN2 4JG, U.K.,
Research keywords: lexicons  

The lexicon; its structure; lexical resources, including published dictionaries; how lexicographers write dictionaries, how they determine what meanings a word has, and how this relates to theoretical discussions of ambiguity and computational work on word sense disambiguation. Language corpora; word frequency distributions; how these vary across language varieties, and how they relate to syntactic and lexical hypotheses; corpus interfaces; automatic and semi-automatic lexical acquisition from corpora. Lexical semantics; formalisms for lexical representation.  

I am currently working on an EPSRC funded project 'Robust Accurate Statistical Parsing' (RASP), which involves integrating and extending several strands of research on robust statistical parsing and automated grammar and lexicon induction, to produce a new parsing toolkit.

For my dissertation I work on a neural network algorithm for conceptual disambiguation and the representation of various natural language constructions. I'm interested in biologically validated NL systems and theories with practical applications. At my paid work, I do rule-based pattern matching.

I like to see systems work.  

Ian Johnson    <ianj@sharp.co.uk> 
Sharp Laboratories of Europe Ltd, Oxford Science Park, Oxford, Oxon., OX4 4GA, U.K.,
Research keywords: machine translation   information extraction   information retrieval  

I manage research and development projects, both in-house and external collaborations. My main areas of interest currently are electronic dictionaries, morphologies, glossing, information extraction/retrieval, machine translation, and applications of language technology within multimedia.  

Geoffrey Leech    <g.leech@lancaster.ac.uk> 
UCREL, Department of Linguistics , Lancaster Univ., Lancs, LA1 4YT, U.K.,

I have been involved in corpus-building and corpus-annotation projects since 1970 (LOB, BNC, etc). Recent projects: improving the grammatical tagging of the BNC; co-editing a book on corpus annotation; doing a large corpus-based frequency grammar of English; parallel tagging using more than one tagger. New project: EAGLES: developing European guidelines for representation and annotation of dialogue. Growing interest in prefabricated chunks of language in NLP.  

Jochen L. Leidner    <leidner@acm.org> 
ICCS, School of Informatics, 2 Buccleuch Place, Edinburgh EH89LW, Scotland, UK.
Research keywords: engineering systems  information extraction  information retrieval 

(1.) Open-domain question answering as answer retrieval (TREC QA and beyond); (2.) Information Extraction methods for shallow, robust event extraction and applications thereof; toponym resolution; application of ML and hybrid methods to IE problems. (3.) internet applications and agents; search engine technology; (4.) the Linguistic Grid, componentware and NLP software engineering (SALE); standardization efforts; (5.) methodology of (and interaction between) the disciplines of Computational Linguistics and Language Engineering.

I investigate MT systems, especially for PC and am interested in their application in the Modern Languages curriculum. I develop CALL courseware and am working on a book on text analysis using the TUSTEP package developed by Tuebingen University.  

Tony McEnery    <mcenery@comp.lancs.ac.uk> 
Lancaster University, Linguistics, Bowland College,, Lancaster, Lancashire, LA1 4YT, U.K.,
Research keywords: pragmatics   text analysis   teaching  

The real keywords for me are ..... corpora, Non-Indigenous Minority Langauges and pronoun resolution. Am mainly interested at the moment in - (in order of importance) - Non-Indigenous Minority Language Engineering and corpus based approaches to pronoun resolution. In both of these areas corpus building, annotation and exploitation are prime concerns for me. I am currently engaged in a project to evaluate the needs/potential of minority language engineering in the UK. Haven't done much with computational pragmatics for a while now, but it certainly kept me awake thinking for a few years in the late eighties/early ninetees.  

Suresh Manandhar    <suresh@cs.york.ac.uk> 
Computer Science, University of York, Heslington, York, YO1 5DD, U.K.,
Research keywords: parsing   grammar induction   syntax  

constraint languages for natural language processing, parsing, typed feature formalisms, machine learning of natural language  

Daniel Mapleson   <d.mapleson@uea.ac.uk>
University of East Anglia
Norwich, Norfolk, NR4 7TJ
Research keywords: dialogue systems   user modelling   pragmatics

I started my doctorate degree at the UEA in the fall of 2001 and am currently studying conversational pattern matching systems. My interest is in trying to take and adapt this type of technology with the goal of improving it at following a dialogue, and recovering from "not understandings" and "misunderstandings" with the user. By studying existing systems such as ALICE, CONVERSE, and commercial products such as the NeuroServer built by NativeMinds, I hope to explore how these systems are so successful in real world domains, and identify what their shortfalls are. Currently, I am developing a system based on ALICE to act as an online agent capable of advising students on their course options, as a platform to explore my ideas further.

I am interested in using corpus-based techniques for IR tasks, and also to apply NLP techniques to corpus data for language analysis.  

Diana Maynard    <d.maynard@dcs.shef.ac.uk> 
Dept of Computer Science, University of Sheffield, Regent Court, 211 Portobello St, Sheffield, S1 4DP, U.K.,
Research keywords: information extraction   named entity recognition   terminology  

My current research is in information extraction, especially named entity recognition and coreference resolution. I am mainly focusing on the problem of adaptive and robust IE, and of adapting existing resources to new domains, languages and applications. I have been involved in the development of the GATE architecture and have written some of them main linguistic components for it. I also have interests in terminology, especially in the fields of medicine and biology, and in the development of annotated corpora and word sense disambiguation.

Interested in most aspects of document management, including annotation and storage of structured documents, information extraction, document retrieval, text summarisation and re-generation. Particularly interested in the use of NLP techniques in any of these.  

Hermann Moisl    <hermann.moisl@ncl.ac.uk> 
Department of English Literature, Language, and Linguistics, Percy Building University of Newcastle Newcastle upon Tyne NE1 7RU
Research keywords: information extraction   semantics   text analysis  

Natural language processing, formal language and automata theory, artificial neural networks, dynamical systems, and multivariate statistical analysis of linguistic corpora. From an earlier life, I also retain an interest and some research activity in the development and cultural role of literacy, and in early Germanic and Celtic languages and history

Roger Mitton    <roger@dcs.bbk.ac.uk> 
Dept Computer Science, Birkbeck College, University of London, Malet Street, London, WC1E 7HX, U.K.,
Research keywords: text analysis   lexicons  

The development of a spellchecker which can provide more help to poor spellers than commercially available spellcheckers currently do, using information about spelling problems, pronunciation, syntactic context, word frequency and collocations. See my web page or my book 'English Spelling and the Computer', published by Longman, 1996.  

Gerald Nelson    <g.nelson@ucl.ac.uk> 
Survey of English Usage, University College London, Gower St, London, WC1E 6BT, U.K.,
Research keywords: syntax   parsing   text analysis  

My main work has been on the International Corpus of English (ICE: http://www.ucl.ac.uk/english-usage/ice/index.htm). I am currently using the parsed ICE-GB corpus in an ESRC-funded empirical study of subordination phenomena in speech and writing.  

Grammar learning (eg categorial grammars, DCGs) in the MDL framework, Controlled Languages for Requirements Engineering, Parsing corpora.  

Nicholas Ostler    <nostler@chibcha.demon.co.uk> 
Foundation for Endangered Languages, 172 Bailbrook Lane, Bath, BA1 7AA, U.K.,
Research keywords: lexicons   semantics   machine translation  

Dr. Chris Paice    <cdp@comp.lancs.ac.uk> 
Lancaster University, Computing Dept, Bailrigg, Lancaster, LA1 4YR, U.K.,
Research keywords: information retrieval   text analysis   information extraction  

Wide interests in information retrieval. Funded research on automatic generation of abstracts of scientific papers, using an information extraction approach. Interests in large-scale text structure, text typology and the meaning of genre.  

The problem any usres faces in user-machine interaction is that he/she has to know a formal language in order to communicate. Natural language generation tries to eliminate part of that problem by producing natural language utterances, on the basis of a given formal input. The Unix Assistant is a research project that attempts to dynamically generate explanatory text from Unix instructions, in Greek. It aims to give on-line explanations supporting users of complex computer systems who are unfamiliar with the formal language necessary to interact with it.  

Research keywords: multimodal systems, information extraction, generation

The focus of my research is the use of context to improve the performance of computer spellcheckers, with particular reference to the needs of dyslexics. The main area which I am investigating is the detection and correction of 'real word' errors (such as there/their) with the ultimate aim of producing a spellchecker that will be able to deal with this type of error.  

I am working on an EU Esprit project called CLIME (Computerized Legal Information Management and Explanation). The aim of the project is to develop software to support access to legal and regulatory information, specifically in the area of maritime law. Together with my colleagues at ITRI I am involved in the application of natural language technology to facilitate the formulation of queries to legal information systems. Furthermore, we are working on the multilingual generation of natural language answers to enquiries in the legal domain.  

Massimo Poesio
University of Essex
Department of Computer Science, Wivenhoe Park, Colchester, CO4 3SQ.
Research keywords: dialogue systems   pragmatics   semantics  

[this is a revised version of my previous entry - some time has passed!] I do work in semantic and pragmatic interpretation, especially on issues related to the development of spoken dialogue systems; I am particularly interested in developing theories supported by psycholinguistic evidence and corpus analysis. The phenomenon I have been studying most in the recent past has been the interpretation of anaphoric expressions, especially pronouns and definite descriptions, with application especially to generation. Because of this work, I have started looking at techniques for the acquisition of lexical knowledge and the acquisition of preferences among semantic interpretation rules. From a theoretical point of view, I have been studying the role of underspecification in linguistic theory and in theories of disambiguation, especially of spoken input.  

Victor Poznanski    <vp@sharp.co.uk> 
Sharp Labs of Europe, Edmund Halley Road, Oxford, Oxon, OX4 4GA, U.K.,
Research keywords: machine translation   information retrieval   information extraction  

Stephen Pulman    <Stephen.Pulman@clg.ox.ac.uk> 
Oxford University
Somerville College Oxford OX2 6HD
Research keywords: semantics   engineering systems   text analysis  

Computational semantics; language and inference; acquiring domain knowledge from text.

Ehud Reiter    <ereiter@csd.abdn.ac.uk> 
University of Aberdeen, Dept of Computing Science, Aberdeen, AB24 3UE, U.K.,
Research keywords: generation   multimodal systems   engineering systems  

I work on NL generation, with a particular emphasis on applied systems-building issues. Currently my main research project is on automatically generating personalised smoking-cessation letters. I am also working on a book on applied NLG, which I hope will appear in 1998 or 1999. I have worked on and off over the years on multimodal (text and graphics) systems, and hope to spend more on time on this in the future.  

We are researching an alternative method of natural language acquisition and generation. The largest, most frequently occurring textual patterns from training corpora containing examples of human dialogue are detected and stored. Appropriate response to a novel utterance is computed by employing fuzzy techniques to detect the subset of textual patterns derived during training that best approximate the novel utterance. Data concerning the infrastructure that approximated textual patterns shared with the utterances from which they were derived are employed to generate the hypothesised response to the novel utterance. The approach does not model conventional language theories, contains no lexicon or grammar rules and is potentially omni-lingual. The research work is shared with Dr.Martin Lefley also of Bournemouth University.  

I am a Senior Research Associate in the Visicast Project. Working on the English-Text-to-Sign-Languages Translation System. My work involves DRS (Discourse Representation Structures) creation as the representation of meaning of English input texts, pronoun resolution and writing HPSG (Head-Driven Phrase Structure Grammar) of British Sign Language for sign generation using HamNoSys (Hamburg Notation System for Sign Languages).

I am a corpus linguist; much of my work is in developing explicit standards for annotating the detailed grammatical structure of real-life spoken and written English (cf. my SUSANNE and CHRISTINE projects). I have also worked on applying stochastic optimization techniques to the parsing problem.  

Quinn Setshedi    <q.setshedi@postgrad.umist.ac.uk> 
University of Manchester
Research keywords: lexicons   machine translation   robust parsing  

Designing a Sesotho Morphological Analyser using Finite State Networks. 

Andrea Setzer    <A.Setzer@dcs.shef.ac.uk> 
University of Sheffield, Department of Computer Science, 211 Portobello Street, Sheffield, South Yorkshire, S1 4DP, U.K.,
Research keywords: information extraction   semantics   text analysis  

My PhD research concentrates on extracting temporal information from newspaper articles (and other genres in the future), representing this information and drawing inferences.  

Mahmoud Shokrollahi-Far  <mah_shok@ucna.ac.ir>
University College of Nabi Akram
Azadi Avenue, P.O.Box 51385-1488, Tabriz, Iran.
Research keywords: machine translation   parsing   engineering systems  

At present I am working on developing a computational functional model for criticising the translations of Holy Qur'an. In January 2002, I am going to start my PhD in Language Engineering at UMIST, Manchester, UK, under the supervision of Prof. Harold Somers.

I don't work directly on linguistics of any sort, but I think some of my work on ontology and representation may be relevant, e.g. my paper on "Actual possibilities" in KR96 and lots of stuff on what a representation is, representations as control states, and generalising notions of syntax, semantics, pragmatics and reasoning. See links in my Web page.  

My research concerns the role of inferences in comprehension. I'm particularly interested in modelling cognitive controls on inference generation - for example, coherence requirements.  

Currently, I am most interested in non-symbolic approaches to Computational Linguistics and Machine Translation, especially the use of language-neutral statistical methods with large corpora for the purposes of knowledge extraction. A particular focus for this research is the development of resources for MT and Computer-Aided Translation involving NIMLs (non-indigenous minority languages). My other research interests include Machine Translation, and the application of Case and Valency grammar to Computational Linguistics. I have recently revived my interest in two topics which I worked on some time ago, namely authorship attribution, with particular focus on the pastiches of well-known authors, and on authors who write for both adults and children. The second topic is possible applications of computational linguistics techniques to language analysis by speech therapists.  

I have worked in a range of areas in language and information processing; my last two projects were on spoken document retrieval; I am currently especially interested in summarisation and in system evaluation.

His research interests cover issues in computational linguistics, artificial intelligence, computer science and cognitive science, including syntax and semantics of natural languages and programming languages, parsing and comprehension of natural language discourse by humans and by machine, natural language generation, and intonation in spoken discourse. Much of his current NLP research is addressed to issues in spoken discourse and dialogue, especially the meaning of intonation and prosody. Some of this work concerns the graphical animation of speaking virtual or simulated autonomous human agents.

Large-scale practical word sense disambiguation, comparing different knowledge sources for semantic disambiguation, deriving lexical information from corpora, shallow parsing evaluation infromation extraction using fast pattern-matching methods automatic construction of extraction dictionaries from corpora  

I have had a long standing interest in Natural Language Engineering, especially application to Information Retrieval, Frame Based approaches toLanguage Understanding including notions of coherence in text and, the use of language resources like existing dictionaries. Recently I have become very interested in the use of Natural Language Engineering for Aids to the Disabled, and more specifically aphasia. I am also heavily involved with applications of Genetic Algorithms, Neural Networks and Case Based Reasoning to applied language processing and other areas.. I lead the University of Sunderland's participation in the EU Framework IV Mable and the EPSRC PSET projects.  

David Tugwell    <davidt@cogsci.ed.ac.uk> 
centre for cognitive science, university of Edinburgh, 2, Buccleuch Place, Edinburgh, EH8 9LW, Scotland,
Research keywords: syntax   robust parsing   parsing  

I am developing a `left-to-right' dynamic model of syntax, mapping strings of words incrementally into semantic representations. Coupled with a suitable control structure, this then forms a parser. I am looking at ways to train such a model from pre-analysed texts to form a probabilistic language model for use in a variety of NLP tasks including speech processing. I am also interested in assessing the plausability of the model as a model of human language comprehension.  

Kees van Deemter    <Kees.van.Deemter@itri.brighton.ac.uk> 
Information Technology Research Institute, University of Brighton, Lewes Road, Brighton, Sussex, BN2 4GJ, U.K.,
Research keywords: multimodal systems   generation   pragmatics  

Having worked in formal semantics/pragmatics and on Concept-to-Speech generation, I am now interested in the question of how pictures, diagrams, and other nontextual elements can be integrated with text to form a coherent multimodal document. Important themes for me are reference to and by parts of a multimodal document and the disambiguating effect of combining different modalities.  

Hy, My name is Alina Viorel and I'm instructor at the Lucian Blaga University of Sibiu, Romania. My theme research in electrical machine area, simuling and modelling using MATLAB-SIMULINK. Any kind of advices will be very useful for me and my students. Thank you  

Example-based approaches to "parsing" and machine translation.  

Nick Webb    <webbnw@essex.ac.uk> 
University of Essex, Wivenhoe Park, Colchester, Essex, CO4 3SQ, U.K.,
Research keywords: dialogue systems   robust parsing   information retrieval  

Currenly working on a engineering Dialogue Systems, but interests also include robust front ends for IR and automatic lexical aquisition techniques for both IR and Dialogue systems.  

My research is in the field of natural language processing, particularly in the development of wide-coverage grammars and efficient parsers. My research in this area is concerned with techniques for compactly encoding and precompiling large lexicalized grammars. I am also interested in statistical approaches to resolving various forms of ambiguity that arise in the analysis of natural languages. My interest in this area includes the development of corpus-based techniques for word sense disambigution, and the use of probabilistic lexicalized grammars to rank parses.  

My research interests are in Artificial Intelligence, in particular Neural Networks, Natural Language Processing and Hybrid Systems. The motivation for this research is twofold: How is it possible to bridge the large gap between real neural networks and high level cognitive language performance? How is it possible to build more effective systems which integrate subsymbolic and symbolic technologies in hybrid systems? Based on this motivation I have directed and worked on several projects, e.g. on hybrid neural/symbolic systems for text processing and speech/language integration.  

Pete Whitelock    <pete@sharp.co.uk> 
Sharp Laboratories of Europe Ltd., Oxford Science Park, Oxford, OX4 4GA, U.K.,
Research keywords: machine translation   lexicons  

My current research is concerned with the development of an English to Japanese 'glosser', i.e. a translator without full TL generation, for Japanese learners to browse English texts and improve their vocabulary skills as they do so. I am interested in a variety of contributory technologies, such as lightweight linguistic analysis (tagging, morphology, shallow parsing), extraction of multi-word expressions from corpora, and sense tagging.  

Yorick Wilks    <yorick@dcs.shef.ac.uk> 
Institute for Language Speech and Hearing, Regenet Court, Portobello St., Sheffield, S.Yorks, S1 4DP, U.K.,
Research keywords: information extraction   dialogue systems   lexicons  

I am interested in, and working on, more things than the three key words: I see both information extraction and conversational/dialogue systems as platforms (for both of which we have prototypes at Sheffield) on or within which a great number of theories and methods can be tested, including the induction of grammars, both for texts and for conversations from dialogue corpora, knowledge and semantic representations, lexicons and methods for adapting them to new domains/senses etc. I would also like to emphasise the role of individual belief and knowledge representations in language understanding, and how they must, to be tractable, be manipulable with some very simple general principles. More recently, I have been researching general engines for content word sense disambiguation in running texts. I have always looked for some kind of generality behind all the areas of activity in NLP--I guess I still believe in the stuff back in my PhD thesis that there is an overall notion of coherence in language structure, one of which syntax is a very weak guide, which somehow minimises the effort required to pull a unique meaning structure for an utterance into being. For me, this principle covers a lot of the phenomena, and I see recent empirical/statistical methods as at last giving data for theories of the general kind I like, and which have been around for a long time in hand-crafted structural forms.  

In Oct 2000 I started a PhD onconfiguring NLG systems to generate texts for adults with poor reading skills. In the past, I have been involved in concept-to-speech, NLG, dialogue modelling and text summarisation.  

I have developed a computational framework, "information compression by multiple alignment, unification and search", intended to integrate such things as parsing and production of natural language, pattern recognition, information retrieval, probabilistic reasoning, unsupervised learning and others. Several of these areas are now quite well developed. The framework supports a novel approach to the representation and use of NL syntax.

Author of MULTICONCORD, a multilingual parallel concordancer, in association with a number of European universities on a LINGUA project. Uses paragraph based non-statistical alignment methodology. Have also written an abridgement porogram based on lexical patterns, and some forensic analysis tools. All for PC/Windows environment.