Description of organization
Xerox Research Centre Europe is part of the Xerox Innovation group made up of 800 researchers and engineers in five world-renowned research and technology centres. The centre creates innovative technologies to support growth in Xerox business process and document management services businesses. Its domains of research stretch from the social sciences to computer science with renowned expertise in machine learning, natural language processing, computer vision, ethnography and services computing. It has won a number of international technology benchmarking competitions and external awards. The centre has been both coordinator and partner in many European Union funded projects and other collaborative projects since its creation in the early 90s. Projects the centre is currently involved in that are relevant to ITFoM include EURECA (Enabling Information re-use by linking clinical Research and care), the PASCAL EU Network of Excellence (Pattern Analysis Statistical Modelling & Computation Learning) and the piloting of a system that assists medical staff in automatically detecting Hospital Acquired Infections from patient records (ALADIN-DTH French Govt. project).
Previous experience
The Machine Learning for Services (MLS) group at XRCE enables the widespread use of predictive analytics in organizations, the adoption of automated evidence-driven decision making, and the emergence of complex adaptive systems. Its research is grounded in statistical data modelling and data-intensive algorithms. Its activities aim to advance machine learning methodology, including generative and discriminative data modelling, structured prediction, collaborative filtering, choice modelling, and relational learning. Its work is articulated along three main research lines:
1. Predictive analytics
2. Mechanism design (or reverse game theory)
3. Knowledge generation
The team has strong expertise in Bayesian inference, as well as convex and nonlinear optimization, and it emphasizes principled ways to deal with uncertainty (noise and partial information). The resulting data-driven solutions capture individual preferences, enable self-learning services and/or borrow from behavioural economics in order to be incentive compatible. They are applied in a wide range of domains, including customer care, healthcare, transportation, and participatory governance. The team is currently involved in the European project FUPOL (www.fupol.eu). Present research collaborations include co-supervision of three PhD students and several Open Innovation projects.
The Enterprise Architecture (EA) research area at XRCE enables the various parts of an organization to articulate information, technology and business strategy in a common decision framework. Its main lines of research are structural pattern recognition methods and technologies, service and business process modelling to bridge domain-specific language and BPM (business process management) generic languages and tooling techniques and interoperability & sustainability to manage information flows across heterogeneous systems and organizations to guarantee the validity and reusability of data and processes over space and time. Some examples of the technology the group has created and made available to the community are the Open Source Domain Specific Language for XML processing (Xeproc) and Circus-DTE, a programming language dedicated to data structure transformation.
Profile of staff members
Cedric Archambeau is research scientist and area manager of the Machine Learning for Services research group. His main research interests include probabilistic machine learning and computational statistics. His work addresses fundamental problems in data assimilation, natural language processing and personalised content creation. In all these domains one of the main challenges is to find efficient ways to deal with today's exponential increase of available digital data and uncertainty. He is also interested in large-scale adaptive systems that combine crowdsourcing and machine intelligence to assist, for example, knowledge workers in their daily tasks. Guillaume Bouchard is a senior research scientist in statistical learning for text understanding and user modelling. He is interested in a wide range of applications including print infrastructure optimization, demand management in transportation or content creation modelling. Shengbo Guo is a research scientist in statistical machine learning and their applications in diverse domains such as recommender systems, discrete choice modelling, dynamic pricing, transportation, and social network analysis.
Jean-Pierre Chanod is senior scientist and manager of the Enterprise Architecture research group. His main research interests relate to natural language understanding, document processing and process modelling. He has been involved in many international research collaborations, in Europe and Asia and contributed to a number of French national projects and EU-funded projects the most recent being the integrated FP6 project VIKEF (Virtual Information & Knowledge Environment) and SHAMAN (Sustaining Heritage Access through Multivalent Archiving) in FP7. Thierry Jacquin is a senior research engineer whose main focus is in design tools and process modelling, based on MDA and Domain Specific Languages (DSL). Thierry was XRCE's technical lead in the VIKEF and SHAMAN projects. Jean-Yves Vion-Dury is a senior scientist who has specialised in programming tools and environments for distributed software components, the theory of formal languages, type systems, logic and process calculi. His current focus is on building practical and theoretical tools based on Semantic Web standards to address semantic interoperability and sustainability issues.
Webpage
Seven recent publications relevant to the project
1) B. Lakshimanarayan, , G. Bouchard, C. Archambeau: Robust Bayesian Matrix Factorisation. Artificial Intelligence and Statistics (AISTATS) 14. JMLR Workshop and Conference Proceedings 15:425-433, 2011.
2) C. Archambeau, M. Opper: Approximate Inference for continuous-time Markov processes. In D. Barber, A. T. Cemgil, and S. Chiappa, Inference and Learning in Dynamic Models. Cambridge University Press, 2011.
3) S. Lise, C. Archambeau, M. Pontil, D. Jones: Prediction of hot spot residues at protein-protein interfaces by combining machine learning and energy-based methods. BMC Bioinformatics, 10: 365-382, 2009.
4) G. Sanguinetti, A. Ruttor, M. Opper, C. Archambeau : Switching Regulatory Models of Cellular Stress Response: Bioinformatics, 25(10): 1280-1286, 2009. Oxford University Press.
5) XML Processing in the Cloud: Large-Scale Digital Preservation in Small Institutions, Peter Wittek, Thierry Jacquin, Jean-Pierre Chanod, Hervé Déjean, Sandor Daranyi, IPDPS 2011 - 25th IEEE International Parallel & Distributed Processing Symposium, Anchorage, (Alaska) USA, May 2011.
6) A generic Calculus of XML editing deltas, Jean-Yves Vion-Dury. DocEng 2011: Proceedings of the 11th ACM symposium on Document engineering, Mountain View CA, USA, September 2011.
7) Xeproc©: a model for XML document processing,available under the Eclipse Public License. http://www.xrce.xerox.com/Xeproc
