People

V&L Net Steering Committee:

Khurshid Ahmad (2010-2013), University College Dublin, IR (SC member)
 
Khurshid Ahmad
My theoretical interests are in neural computing, in terminology and ontology, and in natural language processing, especially information extraction. I am motivated by the way humans deal with information transmitted in different modalities – texts, images, numbers, diagrams. My research in neural computing has led to multi-net neural systems that can mimic the evolution of language and numerosity. Multi-net systems were used in co-locating images and their textual descriptions. My work in terminology and ontology relates to the automatic extraction and deployment of terminology of specialist disciplines in applications as diverse as translation and information extraction; work in ontology has helped develop systems that can track growth of knowledge from early theoretical stages onto patenting and into the market place. The practical objective is to create information systems that can not only deal with information in a number of different modalities but can also learn to deal with different modalities.
Yiannis Aloimonos, University of Maryland, US (Network mentor)
 
Yiannis Aloimonos
The research of Professor Aloimonos is devoted to the principles governing the design and analysis of real-time systems that possess perceptual capabilities, for the purpose of both explaining animal vision and designing seeing machines. Such capabilities have to do with the ability of the system to control its motion and the motion of its parts using visual input (navigation and manipulation) and the ability of the system to break up its environment into a set of categories relevant to its tasks and recognize these categories (categorization and recognition). The work is being done in the framework of Active and Purposive Vision, a paradigm also known as Animate or Behavioral Vision. In simple terms, this approach suggests that Vision has a purpose, a goal. This goal is action; it can be theoretical, practical or aesthetic. When Vision is considered in conjunction with action, it becomes easier. The reason is that the descriptions of space-time that the system needs to derive are not general purpose, but are purposive. This means that these descriptions are good for restricted sets of tasks, such as tasks related to navigation, manipulation and recognition. If Vision is the process of deriving purposive space-time descriptions as opposed to general ones, one is faced with the difficult question of where to start (with which descriptions)? Understanding moving images is a capability shared by all "seeing" biological systems. It was therefore decided to start with descriptions that involve time. Another reason for this is that motion problems are purely geometric and understanding the geometry amounts to solving the problems. This led to a consideration of the problems of navigation. Within navigation, once again, one faces the same question: in which order should navigational capabilities be developed? This led to the development of a synthetic approach, according to which the order of development is related to the complexity of the underlying model. The appropriate starting point is the capability of understanding self-motion. By performing a geometric analysis of motion fields, global patterns of partial aspects of motion fields were found to be associated with particular 3D motion. This gave rise to a series of algorithms for recovering egomotion through pattern matching. The qualitative nature of the algorithms in conjunction with a nature of the well-defined input (the input is the normal flow, i.e. the component of the flow along the gradient of the image) makes the solution stable against noise. Other problems, higher in the hierarchy of navigation, are independent motion detection, estimation of ordinal depth, and learning of space. To illustrate these topics, consider the case of ordinal depth. Traditionally, systems were supposed to estimate depth. Such metric information is too much to expect from systems that are supposed to just navigate successfully. Many tasks can be achieved by using an ordinal depth representation. Such a representation can be extracted without knowledge of the exact image motion or displacement. Recent studies on visual space distortion have triggered a new framework for understanding visual shape. A study of a spectrum of shape representations lying between the projective and Euclidean layers is currently underway. The learning of space can be based on the principle of learning routes. A system knows the space around it if it can successfully visit a set of locations. With more memory available, relationships between the representations of different routes give rise to partial geocentric maps. In hand-eye coordination, the concept of a perceptual kinematic map has been introduced. This is a map from the robot's joints to image features. Currently under investigation is the problem of creating a classification of the singularities of this map. The work on active, anthropomorphic vision led to the study of fixation and the development of TALOS (TALOS), a system that implements dynamic fixation. Since fixation is a principle of Active Vision and fixating observers build representations relative to fixations, it is important to solve fixation in real time and demonstrate it in hardware. TALOS consists of a binocular head/eye system augmented with additional sensors. It is designed to perform fixation as it is moving, in real time. The ideas of Purposive Vision have led to the study of Intelligence as a purposive activity. A four-valued logic is being developed for handling reasoning in a system of interacting purposive agents.
Anja Belz, University of Brighton, UK (Coordinator)
 
Anja Belz
Dr Anja Belz is Reader in Computer Science in the School of Computing, Engineering and Maths at the University of Brighton, where she jointly leads the Natural Language Technology research group. Following earlier research in grammar formalisms, parsing and grammar learning, her research now focuses chiefly on comparative evaluation and probabilistic methods for computational natural language generation (NLG), with a particular interest in application areas that involve both visual and linguistic aspects, such as automatic image description. Belz has held seven major project grants as principal investigator (EPSRC and BA). In the area of probabilistic methods for NLG, Belz and her research team are developing probabilistic language generation methodologies with a particular emphasis on generic input representation. Project outputs include corpora of paired inputs and outputs for training and evaluating data-to-text generation systems, software tools for evaluation, and a wide range of empirical results assessing the performance of different types of systems. Through organising six shared-task evaluation events (with EPSRC support), and numerous publications on the evaluation methodologies, Belz has built up a solid track record in the theory and practice of NLG evaluation. In the EPSRC Network on Vision and Language (V&L Net), Belz and co-investigator Makris (Kingston) are building a forum for researchers from the fields of computer vision and language processing to meet, exchange ideas, expertise and technology, in order to work towards solutions for some of today's most challenging computational problems, including image and video search and description of visual content. Belz has organised and chaired more than a dozen conferences and workshops. She is regularly invited to speak on NLG evaluation. A regular eviewer for the EPSRC, the European Commission (FP7, CIP) and other funding bodies, journals and programme committees, Belz is currently serving as workshops chair for EACL'14, and has served as conference area chair for the areas of document summarisation and language generation at EACL'09, ACL-IJCNLP'09, and ACL'12. Belz is a member of the EPSRC College and the editorial board of Computational Linguistics, the leading journal in natural language processing.
Kalina Bontcheva, University of Sheffield, UK (SC member)
 
Kalina Bontcheva
I am a senior researcher in the Natural Language Processing Group, Department of Computer Science, University of Sheffield. From October 2010, I also hold an EPSRC Career Acceleration Fellowship, working on personalised summarisation of social media. Other topics I've been working on are mining information from patents, sentiment analysis, collaborative environments for text annotation, and GATE. I'm also an area co-chair for Information Extraction at ACL'2010 and workshops co-chair at ESWC'2010 I am the Co-Investigator on the JISC-funded TextVRE project. Between 2006 and 2009 I was the Principal Investigator on 3 EU-funded projects (MUSING, TAO, and ServiceFinder) and the co-ordinator of the TAO consortium, which involved 7 partner institutions. Between 2004 and 2006 I was Sheffield's technical project manager and researcher on the SEKT Integrated Project. Before that, I was Sheffield's technical manager and researcher on the MIAKT e-science project and I also contributed to the AKT project. I have also been working on Sheffield's GATE open-source NLP infrastructure since 1999.
Ted Briscoe, University of Cambridge, UK (SC member)
 
Ted Briscoe
Ted Briscoe has been a member of staff at the Computer Laboratory since 1989, a Reader since 2000 and Professor of Computational Linguistics since 2004. His broad research interests are computational and theoretical linguistics and automated speech and language processing. He directed and was heavily involved in the teaching of the MPhil in Computer Speech, Text and Internet Technology, taught jointly with the Engineering Department. From 1990 until 1996 he was an EPSRC Advanced Research Fellow undertaking research at Macquarie University in Sydney, University of Pennsylvania in Philadelphia and Xerox European Research Centre in Grenoble, as well as at the Computer Laboratory. His specific research interests include (nearly-)deterministic, statistical, and robust parsing techniques, acquiring lexical information from electronic textual corpora and dictionaries, defaults and constraint-based approaches to linguistic description, exploiting prosody and punctuation during parsing, models of human language learning and parsing, and evolutionary simulations of language variation and change. He has published over 70 research articles, edited three books, and been Principal/Co-Investigator or Coordinator of fourteen EU and UK funded projects since 1985. He is joint editor of Computer Speech and Language and on the editorial board of Natural Language Engineering.
Darren Cosker, University of Bath, UK (SC founding member)
 
Darren Cosker
I am currently a Lecturer (UK)/Assistant Prof. (US) at the University of Bath, and also a Royal Society Industry Fellow working with Double Negative Visual Effects (DNeg) in London. The aim of the Royal Society Fellowship with DNeg is to push next generation facial technology for movies. Between 2007 and 2012, I was a Royal Academy of Engineering/EPSRC Research Fellow also at the University of Bath. I am a part of the Centre for Digital Entertainment (CDE) and the Media Technology Research Centre (MTRC), both of which have interests spanning Animation, Visual Interpretation and Digital Effects. I received a BSc in Computer Science with Honours from Cardiff University in 2001, and a Ph.D in Computer Science from Cardiff University in 2006. My thesis explored the synthesis and perceptual analysis of visual speech. I am also a visiting fellow at the Centre for Vision, Speech and Signal Processing (CVSSP) at Surrey University, and visiting researcher at Cardiff University. My research interests lie in the convergence of Computer Vision, Computer Graphics and Psychology - particularly in relation to human motion analysis and synthesis. My overall aim is to create characters that are indistinguishable from real life ones, and to create tools and methods that allow artists to author these characters. I am also interested in applying computer vision and graphics more generally to movie VFX and video games, and have previously worked as an R&D consultant for Double Negative and Sony Computer Entertainment Europe (SCEE).
Roy Davies, Royal Holloway, UK (Network mentor)
 
Roy Davies
Roy Davies is Professor of Machine Vision in the Department of Physics at Royal Holloway, University of London. He has been a leading figure in the areas of image processing and computer vision since the mid-1980s and has contributed greatly to both the field and the community. He was awarded the British Machine Vision Association Distinguished Fellowship in 2005. Roy Davies was educated at Cardiff High School and Jesus College Oxford. He obtained his BA in Physics in 1963 and his DPhil in 1967. The title of his thesis was Electron and Nuclear Resonance Studies in Solids. His early career was based on spin transitions for nuclei and the related electronics, leading to the development of the 'Davies Electron-Nuclear Double Resonance' technique, better known as the 'Davies ENDOR' technique, a method that 35 years later is still regularly referenced in Physics journals. His interest in electronics, noise, and signal extraction or recovery led to the book "Electronics, Noise and Signal Recovery", published in 1993, which integrated the entire area. The book for which Roy Davies is most well known to computer vision students, researchers, and academics all around the world is "Machine Vision: Theory, Algorithms, Practicalities", originally published in 1990 and now in its 3rd Edition. Overall, Roy Davies' career has been involved with making sense of data, and not just images, and in part by negotiation of noise and clutter in a systematic way. His DSc awarded by the University of London in 1996 reflects this major preoccupation. He has published extensively in the leading journals and conference proceedings on many aspects of computer vision, with his work falling under the main headings of image filtering, feature detection, intermediate level analysis, real-time operation, and automated visual inspection. His deep involvement in educational aspects has led to three books and numerous chapters and encyclopaedia articles. Roy Davies' combined interest in image analysis and real-time systems has kept him in popular demand by industry for investigating numerous industrial vision problems. He has specialised in algorithm design for automated visual inspection, and also has interests in vision in less constrained environments such as surveillance, crime detection and prevention, vehicle driver assistance, and laparoscopic surgery. Over many years a large proportion of his research funding has come from grants related to food inspection. This led to his third book, "Image Processing for the Food Industry", published in 2000. Roy Davies is widely recognised for his authority, popularity, and standing in the computer vision community. He has contributed immensely to the computer vision community through his numerous activities in the BMVA and the IEE; his work on many editorial boards (including Pattern Recognition Letters, Real-Time Imaging, Imaging Science and IET Image Processing); and his involvement in education, including examining over 100 PhD students.
Mark Everingham, University of Leeds, UK (SC member, 2010-2012)
 
Mark Everingham
I am a lecturer in the Vision Group of the School of Computing at the University of Leeds. Previously I was an RCUK Academic Fellow at Leeds and before that a Research Fellow in the Visual Geometry Group at the University of Oxford, where I worked on the EU-funded CLASS and CogViSys projects with Andrew Zisserman. I was a Fulford Junior Research Fellow of Somerville College. I completed my PhD in Computer Vision at the Department of Computer Science, University of Bristol, and received my BSc in Computer Science from the University of Manchester. My primary research interests are in high-level understanding of images and video, with an emphasis on probabilistic approaches derived by statistical learning from examples. One application of this work investigated during my PhD has been a mobility aid for the severely visually impaired. My work in this domain has led to applications of probabilistic approaches in other areas such as object extraction from video sequences, and formal methods for evaluating complex vision techniques such as image segmentation algorithms. More recently my research has focused on methods for recognizing objects, people and their pose and behaviour in unconstrained consumer images and video. A central theme to this work has been the exploitation of "weak" but readily-available sources of supervision, with examples including learning models for face recognition from video scripts and subtitles, learning to recognize object categories from textual descriptions alone, learning sign language by "watching TV", and learning articulated pose estimation from inaccurate Mechanical Turk annotation. I am co-organizer of the PASCAL Visual Object Classes (VOC) challenge. I have served as an area chair for the IEEE Conference on Computer Vision and Pattern Recognition (CVPR) and the European Conference on Computer Vision (ECCV), and regularly review for all the main conferences and journals. In 2008 I served as general and program chair of the British Machine Vision Conference (BMVC2008). In 2009 I was elected to the executive committee of the British Machine Vision Association (BMVA). I am an associate editor for IEEE Transactions on Pattern Analysis and Machine Intelligence.
Yannis Goulermas, University of Liverpool, UK (SC founding member)
 
Yannis Goulermas
RESEARCH INTERESTS: * Machine learning / pattern recognition. * Data analysis and visualisation. * Mathematical modelling and optimisation. * Image and signal processing. APPLICATION INTERESTS: * Biomechanics. * Biomedical engineering and imaging. * Bioinformatics. * Text mining. * Market analysis and pricing systems. * Industrial monitoring and control. * Security applications.
Dimitrios Makris, Kingston University, UK (Coordinator)
 
Dimitrios Makris
Dr Dimitrios Makris is a Reader at the Digital Imaging Research Centre at Kingston University. His research interests are Computer Vision, Machine Learning and in particular Motion Analysis and Dimensionality Reduction. His work on learning scene semantic models and on multiple camera surveillance systems has been highly acknowledged by the international research community as reflected by the high number of citations. His recent work is in the area of non-linear dimensionality reduction of time sequences (incl. Best Poster Paper in BMVC11) with applications in human action recognition, style modelling and human body tracking. He is currently the coordinator of the EPSRC Network on Vision and Language (VL-Net). He was the organizer of the IEEE workshops series on Visual Surveillance (2005-2008), the BMVA/EPSRC Summer School on Computer Vision (2009-2010) and the BMVA technical meetings (2008-2012). He was the invited speaker in Second IEEE International Workshop on Tracking Humans for the Evaluation of their Motion in Image Sequences (THEMIS2009).
Katja Markert, University of Leeds, UK (SC member, 2010-2012)
 
Katja Markert
My research interests centre on computational linguistics. More specifically, I am interested in using knowledge-based and statistical techniques for the automatic treatment of problems in semantics, pragmatics and discourse. Recent research and ongoing research includes: * Anaphora Resolution: I currently work on the recognition of information status as well as the resolution of bridging, together with the Heidelberg Institute of Theoretical Studies , where I visited for a year in 2011-2012 as an Alexander von Humboldt fellow>/a>. Previously, in collaboration with Malvina Nissim and Natalia Modjeska,I used simple web queries for bridging the knowledge gap in the resolution in non-pronominal anaphora. We have integrated a web-based feature into a machine learning algorithm for other-anaphora and developed web-based algorithms for the resolution of definite NP coreference as well. * Automatic treatment of regular polysemy and metonymy. In the Mascara project, Malvina Nissim and I explored supervised and unsupervised machine learning techniques for metonymy recognition. We also produced an annotation scheme for metonymies and developed a freely available corpus annotated for metonymies, which is available from my Data page. Our approach for annotation and learning has been adapted for German at the University of Hagen as well as for French. The Business School at the University of Leeds has used our data to research the conceptualisation via metaphor and metonymy of organisations. We also organised a metonymy recognition competition in conjunction with SemEval 2007 in Prague. * I am interested in opinion mining. Together with Fangzhong Su, I developed a system to recognise subjective and objective senses of words in WordNet. Results have been published in Coling 2008, Naacl 2009 and Naacl 2010. The gold standard data we used is available from my Data page. Soon we will make all of WordNet with sentiment annotations available. * My students Andrew McKinlay and Amal Al-Saif are working on discourse relations, the former on the recognition for discourse relations for English, the latter for Arabic. As part of this work we developed the Leeds Arabic Discourse Treebank , the first corpus with annotated discourse relations for Arabic. This work was funded by the British Academy. The Treebank will be available soon via the Linguistic Data Consortium. Google WebDoc: Together with Serge Sharoff and Zhili Wu we work on the automatic recognition of genre for web texts. This work was funded by a Google research award. Past projects: * The recognition of textual entailment and textual inference. Johan Bos and I participated in the Textual Entailment competitions organised in 2005 and 2006 and have developed a system for the automatic recognition of textual entailment that integrates deep and shallow semantic analysis within a machine learning framework. The part of the system performing deep semantic analysis can be downloaded on the Nutcracker website. * Entity Recognition: I was involved in the SEER project at the University of Edinburgh and Stanford University, which has the generalization of current entity recognition tasks as its goal.
Kathy McKeown, Columbia University, US (Network mentor, 2010-2012)
 
Kathy McKeown
Kathleen R. McKeown is the Henry and Gertrude Rothschild Professor of Computer Science at Columbia University. She served as Department Chair from 1998-2003. Her research interests include text summarization, natural language generation, multi-media explanation, digital libraries, concept to speech generation and natural language interfaces. McKeown received the Ph.D. in Computer Science from the University of Pennsylvania in 1982 and has been at Columbia since then. In 1985 she received a National Science Foundation Presidential Young Investigator Award, in 1991 she received a National Science Foundation Faculty Award for Women, in 1994 was selected as a AAAI Fellow, and in 2003 was elected as an ACM Fellow. McKeown is also quite active nationally. She serves as a board member of the Computing Research Association and serves as secretary of the board. She served as President of the Association of Computational Linguistics in 1992, Vice President in 1991, and Secretary Treasurer for 1995-1997. She has served on the Executive Council of the Association for Artificial Intelligence and was co-program chair of their annual conference in 1991. Kathy's interests lie in the area of natural language processing and in particular, natural language generation. Kathy and her group are currently working in three main areas. The first, summary generation, involves the generation of natural language text, or summaries, from data such as stock market statistics. The group also have a number of projects on text summarization. Their work includes generating summary updates over live multimedia information, domain independent generation of summaries using a combination of statistical and linguistic techniques, and generation of summaries across multiple, medical articles. In the second, statistical natural language, the group are using statistical analysis of large text corpora to identify constraints on how words are used. Such results can be used to automate the development of a lexicon, or dictionary. Finally, the group are also working on the generation of multimedia explanation, developing techniques to coordinate language and graphics, to produce explanations in the context of a human computer interface for medical information. Part of the interest here is on the generation of spoken language and differences with generation of text.
Frank Keller, University of Edinburgh, UK (SC founding member)
 
Frank Keller
I am a reader in the School of Informatics at the University of Edinburgh. I am affiliated with the Probabilistic Models of Language Group and with the Language and Vision Group. My research focuses on how people solve complex tasks such as understanding language or processing visual information. My approach to understanding human cognition combines experimental techniques such as eyetracking with computational modeling. Eyetracking makes it possible to build up a highly accurate picture of where people look when they read a sentence, speak a word, or view a visual scene. The data generated by eyetracking experiments allows us to build computational models that simulate the behavior we want to study. Such models predict, for instance, which words humans fixate when they read a text, or which objects they focus on when searching a visual scene.
William Smith, University of York, UK (SC founding member)
 
William Smith
My research interests are related to face processing, shape-from-shading and reflectance modelling. These include: * Estimating 3D shape from single images * Face shape and appearance models * Illumination and reflectance modelling (I am particularly interested in applying biophysical models of skin reflectance to shape estimation and recognition tasks) * Face recognition under extremes of illumination, pose and expression * Statistical shape modelling (particularly for directional data or data lying on complex manifolds) * The psychology and neuropsychology of face perception and shape-from-X * Modelling craniofacial variation
Yorick Wilks, University of Sheffield, UK (Network mentor)
 
Yorick Wilks
Yorick Wilks is Professor of Artificial Intelligence at the University of Sheffield, and is also a Senior Research Fellow at the Oxford Internet Institute at Balliol College. He studied math and philosophy at Cambridge, was a researcher at Stanford AI Laboratory, and then Professor of Computer Science and Linguistics at the University of Essex, before moving back to the US for ten years to run a successful and self-funded AI laboratory in New Mexico, the Computing Research Laboratory, a new institute set up by the state of New Mexico as a center of excellence in artificial intelligence in 1985. His own research group in New Mexico was rated among the top five in the US in its area by the Laboratory's International Advisory Board, and it became totally self-supporting with grants by 1990. In 1993 he took up a chair of Artificial Intelligence at the University of Sheffield, and also became founding Director of the Institute of Language, Speech and Hearing (ILASH). Since then he has raised over $50 million in grants from UK research councils and the EC since 1993, and the Sheffield Natural Language Processing Research Group constitutes a major UK group in the area. He has participated in and been the PI of numerous UK, US and EC grants, including the UK-government funded Interdisciplinary Research Centre AKT (2000-2006) on active knowledge structures on the web (www. aktors.org). He has published numerous articles and nine books in that area of artificial intelligence, among which are Electric Words: dictionaries, computers and meanings ( 1996 with Brian Slator and Louise Guthrie) from MIT Press, and Machine Translation: its scope and limits, in 2008 from Springer. His most recent book is Close Encounters with Artificial Companions ( Benjamins, in press 2010). He is a Fellow of the American and European Associations for Artificial Intelligence, a member of the UK Computing Research Council and on the boards of some fifteen AI-related journals. He designed CONVERSE, the dialogue system that won the Loebner prize in New York in 1997, and was the founding Coordinator of the EU 6th Framework integrated project COMPANIONS (4 years, 15 sites, 13meuro) on conversational assistants as personalised and permanent web interfaces. The distinguishing feature of a Companion is its detailed knowledge of its owner as well as the wider-world; its current major implementation is as a elicitor and organizer of someone's personal knowledge and digital records, but the general concept is being adapted to learning, health and travel environments. In 2008 he was awarded the Zampolli Prize at LREC-08 in Marrakech, and the ACL Lifetime Achievement Award at ACL08 in Columbus, OH. In 2009 he was awarded the Lovelace Medal by the British Computer Society. In 2009 he was elected a Fellow of the ACM.

 

Technical support:

Gary Brooks, System Administrator, University of Brighton
 

Administrative support:

Rebecca Tonge, Project Administrator, University of Brighton
 

Public relations and media liaison:

Phil Mills, Communications Officer, University of Brighton