Curriculum Vitae - Roger Evans

Note: This document is an abridged version of my full CV, in which some details and older information are omitted. My full CV, as well as this version and a one page summary, are available on-line at http://www.brighton.ac.uk/nltg/home/Roger.Evans/private/cv. If you are reading this on-line, the links in this document will take you to the corresponding more detailed sections of the full version.


O V E R V I E W

Name Roger Paul Evans
Date of birth 13 August 1959
Qualifications D.Phil in Cognitive Studies (Computational Linguistics), Sussex, 1987
Certificate of Advanced Study in Mathematics, Cambridge, 1981
BSc Hons (1st class) in Mathematics, Warwick, 1980
Current post Reader in Computer Science at the University of Brighton.
Research interests Foundations, architectures and applications of computational linguistics and language engineering: particularly lexical description and representation; natural language generation architectures; control issues for wide-coverage generation; statistical and symbolic methodologies; multilingual issues; eScience and GRID applications for NLP.
Current responsibilities

Leader of the Natural Language Technology Group in the School of Computing, Mathematics and Information Sciences (CMIS), University of Brighton.
Researcher and research manager in natural language systems and data representation, natural language generation and lexical representation.
Principal investigator for EPSRC project COGENT.
PhD supervisor and examiner.
Member of the School Research Strategy Group and the Faculty Research Committee.

Additional Information Member of the EPSRC Peer Review College, 2006-2009.
Holder of an SERC Advanced Fellowship, 1988-1994.
Deputy Head of ITRI, 1994-2000.
54 publications, 26 reports and 100 conference and other presentations.
Secured 14 UK and European grants since 1988, with total income of over £1.5 million.
Co-developer of the lexical representation language DATR, the WYSIWYM knowledge editing technique, and the RAGS generation architecture. First European participant in ARPA MUC evaluations. Chair of organising committee for ECAI-98 and INLG-04.
Contact details Natural Language Technology Group
University of Brighton
Lewes Rd, Brighton
BN2 4GJ, UK

Phone: +44 1273 642907
Fax: +44 1273 642908
Email: R.P.Evans@brighton.ac.uk
URL: http://www.brighton.ac.uk/nltg/home/Roger.Evans

CV: http://www.brighton.ac.uk/nlth/home/Roger.Evans/private/cv

A C A D E M I C    E M P L O Y M E N T

1993-present University of Brighton.
Reader in Computer Science (1997-present).
School of Computing, Mathematical and Information Sciences (2005-present)
Leader of the Natural Language Technology Group
Responsible for all aspects of group activities and development: research strategy, funding, mananagement and own research, student and staff recruitment, supervision, commercial activities and intellectual property issues and technical infrastructure. Principal investigator on the COGENT project. For full details see "Current Activities."
Information Technology Research Institute (ITRI) (1993-2005)
Leader of research group on lexicons and corpus-based lexicography.
Deputy Head of ITRI (1994-2000)
Principal research fellow (1994-1997)
Senior research fellow (1993-1994)
1984-1993 Cognitive and Computing Sciences (COGS), University of Sussex.
SERC Advanced Fellow, (1988-1994) initially in COGS (suspended 1989-1990), and transferred to ITRI in 1993.
Research fellow (Grade II) (1989-1990)
Research fellow (1984-1987)


P R O F E S S I O N A L    A C T I V I T I E S

Teaching and examining

Research supervisor, University of Sussex and University of Brighton.
Supervised four D.Phil/PhD students to successful completion (no failures). Also supervised several MSc dissertation projects and several students on short term placements. Currently supervising four PhD students at Brighton on the topics of orthography and phonology, narrative style, characterisation of genre and generic topic identification. Member of thesis committee for one DPhil student at Sussex on the topic of diathesis alternations.

PhD examiner: Brunel, Durham, Edinburgh, Brighton, Sussex.
MSc external examiner: Manchester.
BSc course external examiner: Ulster.
Thesis proposal assessor: Aberdeen.

Guest lecturer in Topics in Language Technology undergraduate module, and MSc in Lexical Computing and Lexicography.

Experience of course design and validation, assessment, moderation and exam board processes, plagiarism issues.

Refereeing and monitoring

Research proposal referee for EPSRC, BBSRC, ESRC, JCI, SSHRC (Canada). Academic referee for Computional Linguistics, Journal of Linguistics, Journal of Natural Language Engineering, Academic Press, MIT Press. Conference referee for ECAI-98, ACL-98, EACL-99, ECAI-00, SBIA-00, ACL-01, ECAI-02, EACL-03, ACL-04, INLG-04, AHM-05, EACL-06.

Conferences and workshops

Co-organiser of INLG-04. Chair of organising committee for ECAI-98.
Co-organiser of the AHM2005 Mini-Workshop on Text Mining for eScience, 2005. Organising committee member of the EACL-03 Slavic Morphology Workshop, 2003;ACL-SIGLEX SENSEVAL-2 word sense disambiguation workshop, 2001; AISB-99 workshop Reference Architectures and Data Models, 1999; ACL-SIGLEX SENSEVAL word sense disambiguation workshop, 1998. Co-organiser of the ECAI-98 workshop Multilinguality in the Lexicon II , 1998.

Professional bodies

Member of the Association for Computational Linguistics (ACL) since 1982, Computational Linguistics UK (CLUK) group, and SIGGEN (ACL Special interest group for natural language generation).
Member of Association of University Teachers.

Management

Member of the Research Strategy Group of the School of Computing Mathematical and Information Sciences, 2005-present.
Member of the Faculty Research Committee of the Faculty of Management and Information Sciences, 2005-present.
Member of the ITRI Strategy Management Group, participating in the management and development of overall research, teaching and recruitment strategy, 2000-2005
Member of the Faculty Academic Board for the Faculty of Science and Engineering.
IT Faculty representative on Faculty of Health Graduate Medical School Board, 2000-2002.
Member of the University of Brighton Innovation Strategy and Support Group, 1995-2000.
Member of judging panel for the University of Brighton Innovation Awards, 2004.

Training Managing Research Students (1/2 day course), Brighton 2003.
Staff Development Reviewing (1/2 day course), Brighton, 2001.
Time management (1 day course), Brighton, 1998.
Staff recruitment (1 day course), University of Brighton, 1997.
Research supervision (2 day course), University of Brighton, 1995.
Staff appraisal (1 day course), University of Sussex, 1992.


R E C E N T    P U B L I C A T I O N S     A N D     R E P O R T S

Note: Co-authorship indicates genuine collaboration in all cases.
Books Belz, A., R. Evans and P. Piwek, 2004, Natural Language Generation - Third International Conference, INLG 2004 Springer, Heidelberg, Germany.

Journal articles Evans, R., P. Piwek, L. Cahill and N. Tipper, (to appear) "Natural Language Processing in CLIME -- a multilingual legal advisory system," Natural Language Engineering

Mellish, C. and R. Evans, 2004 "Implementation Architectures for Natural Language Generation," Natural Language Engineering 10(3/4), pp. 261-282.

Mellish, M., M. Reape, L. Cahill, R. Evans, D. Paiva, D. Scott, 2004 "Towards a Reference Architecture for Generation Systems," Natural Language Engineering 10(3/4), pp. 1-25.

Evans, R., R. Gaizauskas, L.J. Cahill, J. Walker, J. Richardson and A. Dixon, 1996, "POETIC: A System for Gathering and Disseminating Traffic Information," Natural Language Engineering 1(4), pp. 1-25.

Evans, R. and G. Gazdar, 1996, "DATR: a Language for Lexical Knowledge Representation," Computational Linguistics , 22(2), pp. 167-216.

Book chapters Paiva, D. S. and R. Evans, 2004, "A framework for stylistically controlled generation," in Belz, A., R. Evans and P. Piwek, Natural Language Generation - Third International Conference, INLG 2004 Springer, Heidelberg, Germany, pp. 120-129.

Evans, R., G. Gazdar and D. Weir, 2000, " 'Lexical Rules' are just lexical rules," in A. Abeille and O. Rambow (eds.) Tree adjoining grammars: linguistic, formal and computational properties, CSLI Lecture Notes, Chicago University Press, pp. 71-100.

Conference papers

Baud, R. H., M. Nyström, L. Borin, R. Evans, S. Schulz, P. Zweigenbaum, 2005 "Interchanging Lexical Information for a Multilingual Dictionary," in C.P Friedman, J. Ash and P. Tarczy-Hornoch (eds.), Biomedical and Health Informatics: From Foundations to Applications to Policy, Washingon DC, USA, pp. 31-35.

Carroll, J., R. Evans and E. Klein, 2005 "Supporting text mining for e-Science: the challenges for Grid-enabled Natural Language Processing," in S. Cox and D.W. Walker (eds.), Proceedings of the UK e-Science All Hands Meeting 2005, Nottingham, UK, ISBN 1-904425-53-4, pp. 233-238.

Paiva, D. S. and R. Evans, 2005, "Empirically-based control of natural language generation," in K. Knight, H.T. Ng and K. Oflazer (eds.), Proceedings of the 43rd Annual Meeting of the Association for Computational Linguistics (ACL 2005), Ann Arbor, USA, pp. 58-65.

Erjavec, T., R. Evans, N. Ide and A. Kigarriff, 2003 "From Machine Readable Dictionaries to lexical Databases: the CONCEDE experience," Proceedings of COMPLEX 2003, Budapest, Hungary.

Evans, R., C. Tiberius, D. Brown, and G.G. Corbett, 2003 "A large-scale inheritance-based morphological lexicon for Russian," Proceedings of the EACL-03 workshop on Slavic Morphology, Budapest, Hungary.

Koeling, R., A. Kilgarriff, D. Tugwell and R. Evans, 2003 "An evaluation of a lexicographer's workbench: building lexicons for machine translation," Proceedings of the EACL-03 EAMT workshop, Budapest, Hungary.

Evans. R., P. Piwek and L.J. Cahill, 2002 "What is NLG?" Proceedings of INLG 2002, New York, USA.

Cahill, L., J. Carroll, R. Evans, D. Paiva, R. Power, D.Scott and K. van Deemter, 2001 "From RAGS to RICHES: exploiting the potential of a flexible generation architecture" Proceedings of the 39th Annual Meeting of the Association for Computational Linguistics (ACL 2001), Toulouse, France

Cahill, L., C. Doran, R. Evans, C. Mellish, D. Paiva, M. Reape, D. Scott and N. Tipper, 2000 "Reinterpretation of an existing NLG system in a generic generation architecture," Proceedings of INLG 2000 Tel Aviv, Israel.

Cahill, L., C. Doran, R. Evans, R. Kibble, C. Mellish, D. Paiva, M. Reape, D. Scott and N. Tipper, 2000 "Enabling resource sharing in language generation: an abstract reference architecture," Proceedings of LREC 2000 Athens, Greece.

Mellish, C., R. Evans, L. Cahill, C. Doran, D. Paiva, M. Reape, D. Scott and N. Tipper, 2000 "A representation for complex and evolving data dependencies in Generation," Proceedings of ANLP-2000 pp. 119-126., Seatle, USA.

Piwek, P., R. Evans, L. Cahill, and N. Tipper, 2000 "Natural Language Generation in the MILE System," Proceedings of the IMPACTS in NLG Workshop pp. 33-42., Schloss Dagstuhl, Germany

Tiberius, C. and R. Evans, 2000 "Phonological feature based Multilingual Lexical Description," Proceedings of TALN 2000, Geneva, Switzerland

Conference posters and demos

Généreux, M. and R. Evans 2006 "Distinguishing affective states in weblogs," Proceedings of the AAAI-2006 Spring Symposium on "Computational Approaches to Analyzing Weblogs", Stanford, California, USA.

Power, R. and R. Evans, 2004 "WYSIWYM with wider coverage," Research notes and demonstration papers at ACL-04, Barcelona, Spain.

Power, R. and R. Evans, 2004 "WYSIWYM with wider coverage," INLG-04 poster session, New Forest, UK.

Evans, R., K. van Deemter, A. Belz, J. Teeple, D. Weir, J. Carroll, D. Paiva, E. Esteve Ferrer, 2004 "Controlling wide-coverage generation - the COGENT project," INLG-04 poster session, New Forest, UK.

Evans, R. and R. Power, 2003 "WYSIWYM - building user interfaces with natural language feedback," Research notes and demonstration papers at EACL-03, pp. 203-206, Budapest, Hungary.

Kilgarriff, A., R.Evans, R. Koeling, M. Rundell and D. Tugwell, 2003 "WASPBENCH: a lexicographer's workbench supporting state-of-the-art word sense disambiguation," Research notes and demonstration papers at EACL-03, Budapest, Hungary.

Erjavec. T, R. Evans, N. Ide and A. Kilgarriff, 2000 "The CONCEDE Model of Lexical Databases," Proceedings of LREC 2000 Athens, Greece.

Other reports

Belz, A., R. Evans and P. Piwek 2004 INLG04 Posters: Extended abstracts of posters presented at the Third International Conference on Natural Language Generation ITRI technical report ITRI-04-01, ITRI, University of Brighton.

Evans, R., C. Tiberius, D. Brown, and G.G. Corbett, 2003 Russian lemmatisation with DATR, ITRI technical report ITRI-03-23, ITRI, University of Brighton.


R E C E N T   R E S E A R C H    P R O J E C T S

Note: I was principal investigator on all projects except where otherwise stated.
COGENT
2003-2006
Controlled generation of text
EPSRC: £208,434 (1RF + 1 student for 36 months).
Investigation of non-determinism and ambiguity in wide-coverage generation (joint with Sussex).
Further information: http://www.itri.brighton.ac.uk/projects/cogent
HALO/DarkMatter
2004-2005
Project HALO - DarkMatter consortium
Principal investigator: Richard Power.
Development of a 'Digital Aristole' advanced question-answering system, funded by Vulcan Inc.
Further information: http://www.projecthalo.com
Semantic Mining
2004-2005
Semantic interoperability and data mining in Biomedicine (Network of Excellence)
EU award to Donia Scott.
Development of generic methods and tools supporting medical and biomedical informatics.
Further information: http://www.itri.brighton.ac.uk/projects/semanticmining
EUROMAP
2001-2003
EUROMAP/HOPE 2001
EU IST initiative, 112,379 EURO(1 RF for 19 months).
Promoting greater awareness and fast take-up of HLT in Europe (UK node coordinator).
Further information: http://www.itri.brighton.ac.uk/projects/euromap
PILLS
2001
Pharmaceutical Instructions Language Localisation System.
EU award to Donia Scott and Richard Power.
Development of a multilingual authoring tool for pharmaceutical information.
Further information: http://www.itri.brighton.ac.uk/projects/pills
MATS
2000-2001
Manual Tagging for SENSEVAL
EPSRC: £15,341 (14 months).
Lexicography support for the second SENSEVAL workshop.
Further information: http://www.itri.brighton.ac.uk/events/senseval
WASPS
1999-2002
A semi-automatic lexicographer's workbench for writing word sense profiles.
EPSRC: £287,207 (2 RFs for 3 years).
Development of a lexicographer's workbench.
Further information: http://www.itri.brighton.ac.uk/projects/wasps
GREG
1999-2001
A Georgian, English, Russian and German multilingual valency lexicon for natural language processing.
EU INCO-Georgia: 5,000 EURO (5pm over 2 years).
Development of a multilingual valency lexicon and lexicon development framework.
Further information: http://www.itri.brighton.ac.uk/projects/greg
CLIME
1998-2001
Computerised legal information management and explanation.
EU ESPRIT: 426,000 EURO (2 RFs for 3 years).
Development of a web-based legal advisory system for shipping regulations.
Further information: http://www.bmtech.co.uk/clime/index.html
RAGS
1998-2001
RAGS: Reference Architecture for Generation Systems.
EPSRC award to Donia Scott.
Development of a standard architectural framework for NLG systems (joint with Edinburgh).
Further information: http://www.itri.brighton.ac.uk/projects/rags
CONCEDE
1998-2000
Consortium for Central European Dictionary Encoding
EU INCO-COPERNICUS: 51,000 EURO (15pm over 2.5 years).
Development of medium-sized electronic dictionaries for 6 central European languages (overall project coordinator).
Further information: http://www.itri.brighton.ac.uk/projects/concede


R E S E A R C H     I N T E R E S T S

I am currently a Reader in Computer Science and leader of the Natural Language Technology Group, in the School of Computing, Mathematical and Information Sciences, University of Brighton, having pursued a full-time research career since completing my DPhil in 1987. After studying mathematics at Warwick and Cambridge, with interests in logic, abstract algebra and category theory, plus a range of computer science options, my doctoral training at Sussex (1981-1986) introduced me to Artificial Intelligence in general, and Computational Linguistics/Natural Language Processing in particular. This was followed by a research post at Sussex (1984-1988) and five years as an SERC (now EPSRC) Advanced Fellow (1988-1993). In 1993 I moved to the University of Brighton to join the Information Technology Research Institute (1993-2005), and became leader of the Natural Language Technology Group in 2005.

My research is situated very much at the 'computational' end of Computational Linguistics, with a focus on processing and data architectures, algorithms, representation formalisms and research methodology, as well as an interest in developing practical applications, user interfaces and tools. I have pursued these interests in a range of application areas within Computational Linguistics and Human Computer Interfaces, including natural language understanding and generation, text mining, question-answering natural language generation, lexical representation and computational lexicography. These application areas are of interest in their own right, but also provide challenging problems for computer science technology more generally. As well as addressing language processing problems directly, I am interested in applying insights gained from their study to a wider class of computational scenarios, such as complex architectures supporting intelligent systems.

Controlling generation
I am currently a co-principal investigator in a major generation project, COGENT (EPSRC, joint with Sussex), which aims to investigate systematically the issues that arise with wide-coverage language generators. In particular, we intend to establish a large (100,000 entry) corpus of semantic inputs compatible with the LinGO wide-coverage generator, from which the generator will be able to produce a much larger number of textual realisations. Using this framework the project will explore non-determinism (generating multiple texts from a single input) and ambiguity (generating a single text from multiple inputs) in order to develop techniques to allow the generator to be effectively controlled. This represents a step forward both in wide-coverage generation research itself, and in the methodology for undertaking research in this area.

Lexical representation
My work on lexical representation in DATR continues in a number of directions:
  1. with Gerald Gazdar I am investigating incorporation of probabilistic information into DATR descriptions, and exploring how to learn inheritance-based statistical models. We are also developing an XML syntax for DATR, in part with a view to facilitating the use of DATR's descriptive capabilities in the development of ontological and ontolexical resources, for example in the Semantic Web.
  2. I am exploring the use of DATR to represent 'extended lexicalism', that is lexical items in context, rather than as just isolated words. As an initial experiment I have developed a view of the part-of-speech tagging problem as a lexical description problem, and used probabilistic DATR to construct a simple tagger.
  3. I am working with colleagues at Surrey who are developing morphological analysis software for morphologically complex languages such as Russian.
Architectures for NLP and intelligent systems
NL generation, although still rather a niche area of NLP, introduces some important different considerations often ignored by the NL understanding community, but relevant more widely to complex and intelligent systems in general. The key difference is that while NL parsing can approach the language processing problem in a layered way (working up through lexical structure, syntactic structure, semantic structure, conceptual structure, and simply stopping when the problems get too difficult), NLG does not have this luxury. It has to address the complexity of linguistic data and data interactions from the outset, and address the difficult question of 'input' - the boundary between linguistic and non-linguistic processing. This makes it an interesting application area for some challenging issues:
  1. What kind of architectures, especially data architectures are required to support the complex needs of NLG? We began to address this question in the RAGS project, and I am now exploring a more abstract version of this framework.
  2. One interpretation of this approach of particular interest is of data manipulation as model-satisfaction/building in a logical sense. This view is suggestive of a programming paradigm based entirely on logical model-building - 'model-based programming'.
  3. What are the relationships between the kind of 'event-driven' process architectures we used in RAGS and algorithms used elsewhere in NLP (eg chart parsers and generators) and other more general dynamic programming-style approaches?
  4. The data representations required to support these architectures are complex. Can this complexity can be mitigated by use of inheritance, statistics, or a mix of the two?

Large-scale probabilistic language analysis
Current work on NL parsing is dominated by statistical and corpus-based approaches and consequently still very 'shallow' - deeper processing requires more data, more processing power etc.. Several responses to this are of interest to me:
  1. Analysis of the underlying mathematics required to support this kind of probabilistic approach - in particular I am beginning to question whether some of the fundamental assumptions of probability theory are the best approach to this kind of computational modelling of language.
  2. Integration of statistical and inheritance-based approaches - in particular, the combination of default inheritance logic representations, extended lexicalism and probabilities as a powerful way of addressing sparse data issues. Applications in areas such as multi-word expressions, term extraction and deeper syntactic and semantic analysis.
  3. The implications of extended lexicalist models for theories of grammar and notions of grammaticality.
  4. Use of Grid technology, to increase the scale of computation possible, and improve the methodology and replicability of results.

Lexical and ontological resources
Through my previous work on lexical resources in a range of contexts, I have a good appreciation of the requirements for effective practical architectures to support reusable large-scale lexicons, in particular multilingual lexicons covering a wide range of lexical phenomena (phonology, morphology, syntax semantics etc.). It is also evident that much current work on large-scale ontologies (for example for the Semantic Web) faces representation and practical issues with many similar features. I am keen to develop a more integrated picture of 'ontolexical' resources which combines the strengths of both fields and can leverage the development of resources between the two.

Document management
Work in text mining, symbolic authoring, document structure and computational lexicography inform my ongoing interest in document content and management issues, ranging from relatively mundane document management (email management, spam detection etc.), to complex XML-based inheritance representations for structured documents (such as dictionaries), and 'dynamic' documents - single representations which present themselves differently according to context (possibly using NLG technology).


O T H E R    P R O F E S S I O N A L     A C T I V I T I E S

Management

I have always engaged in management and infrastructure activities, as a benefit both to my own personal development and to the general health of the research teams I have worked in. As a manager of research and technical staff (group size ranging from 3 to about 7) I place great importance on the developmental needs of my staff: security of funding, opportunities for career development and promotion, equality of opportunity. I have also participated in more general management of the institute, managing and budgeting the computing facilities and technical staff, coordinating technical and estates aspects of a move to new premises, As Deputy Head of ITRI for six years (the post rotated to a colleague in 2000), and now as Leader of the NLTG, I have a considerable experience of strategic planning, budgeting and recruitment.

Innovation and commercial liaison

I advise on intellectual property issues and exploitation within the research group and the wider university, and coordinate commercial liaison activities. I was project manager for the EUROMAP project (EU) which established the UK node of the EUROMAP network, seeking to promote take-up of human language technologies in commerce and industry.

At the University level, I served on the Innovation Strategy Support Group, a directorate subcommittee to develop the University's innovation and commercialization activity, which managed pump-priming funding for new innovations, coordinated the the production of an IPR code of practice for staff and the establishment of the Business Services unit to service and develop commercial activity. In 2004 I was invited to join the judging panel for the University's annual Innovation Awards, which offers prizes for the best innovation and best business plan from staff and students at the University.

Research community

I am active in the 'infrastructure' of the wider research community. During the 1990's I spent six years as treasurer of AISB, during which time I led a competitive bid for, and subsequently chaired, ECAI-98, the main European AI conference, which ran very successfully in Brighton in August 1998. More recently I have been involved with various workshops and evaluation initiatives, and I was lead organiser for INLG-04, the main international conference on natural language generation, held in July 2004. I also review regularly for journals, conferences and research councils.


O T H E R    I N F O R M A T I O N

Computing skills I am an experienced computer user at all levels - system/network manager, system programmer, application programmer and end-user - with industrial and academic experience of a wide range of programming styles and techniques. I am currently an expert programmer in Java, Prolog and Pop11, have reasonable proficiency in Perl, Python, Javascript, shell scripting etc., and memories of using C, LISP, Pascal, Fortran, Basic etc. I am comfortable with procedural, declarative and object-oriented programming styles with experience ranging from assembler to 4GLs.

I have a good understanding of the Unix operating system (in particular Solaris, and more recently linux), and detailed knowledge of the X Window system. I also have a good understanding of the internet (protocols, services, configuration issues), email (SMTP, MTA's MUA's, IMAP, sendmail, Exim etc.) world-wide-web technology (servers, web pages, client browsers etc.), and mark-up technology (HTML, XML, SGML etc.). I am a competent user of Windows and MS Office applications, currently developing greater familiarity with system level issues in Windows 2000/XP.

Outside interests Much of my free time is taken up with my family (wife Linda and children Matthew, Benjamin, Alexander and Jack), and I enjoy sport (especially badminton and soccer - the latter mainly as a spectator these days), music (I play piano, violin, classical guitar and saxophone, but not particularly well) and bridge. I am a former Governors at St. Luke's Infants School, Brighton (former chair and vice-chair, with specific responsibility for finance, personnel and ICT), and treasurer of Brighton Cougars Basketball Club.

28 January 2006