 |
 |
Could Computational Statistics in the UK benefit from a Collaborative Computational Project (http://www.ccp.ac.uk) or a similar undertaking? This is the question we propose to examine during a free agenda-setting workshop on the 28th of July in Manchester. CCPs are a unique British model for collaboration and development of computationally intensive codes supporting scientific communities such as physicists, chemists and others with considerable success. CCPs have been particularly successful in bringing together key scientific expertise, resources and development for codes for large-scale numerical applications. These shared developments have made significant scientific progress possible and thereby raised the profile of UK research and industry in these areas worldwide. In general CCPs: 1. Carry out flagship code development projects 2. Maintain and distribute code libraries 3. Organise training in the use of codes (sometimes in the form of annual schools) 4. Hold meetings and workshops 5. Invite overseas researchers for lecture tours and collaborative visits Large scale computational statistical methods are used in many areas, e.g. social science, meteorology, pharmaceuticals and finance. Is there any common ground in the computational methods/ science of the different areas? Is there any cross fertilization between application areas and methodologies that could occur? If any of the answers contain a yes then computational statistics could benefit from a CCP. It is our impression that the development of computational science for statistics has been very fragmented and that a lot of the methods that have been deployed are very inefficient or suboptimal for the computational systems on which they are used. We have invited a range of interested computer scientists and statisticians to this meeting in order to examine the opportunities of a CCP. Following a series of short presentations and a facilitated brainstorming process we are hoping for statement of the landscape and the beginning of an agenda on how to progress, develop and collaborate. We would like to invite you to participate in this workshop and get your help in shaping the future of code developments for Computational Statistics. Schedule: Venue: Room G.035/036 Arthur Lewis Building, University of Manchester. 10:30-11:00. Registration and Coffee. 11:00-13:00: * Introduction and Aims of the Meeting. Peter Halfpenny. * Data Driven Methods and the Requirement for HPCs. David Hand. * Putting Statistical Science and Computational Science Together. John Logsdon. * Some Limitations of Statistical Software for Modelling Social Processes. Rob Crouchley. * HPC, Hartree Centre and the Grid. Rob Allan. 13:00-14:00. * Lunch. 14:00-15:30. * Articulating the Agenda. All. * Strategy Formulation and Next Steps. All. 15:30-16:00. * Tea and Conclusions. If you would like to attend the workshop, please register using the on-line system at http://conf.dl.ac.uk/event.asp?eventID=145 For additional administrative information please contact Teresa Wisniewska ([log in to unmask]) at Lancaster – 01524 592175. -- Best wishes John John Logsdon "Try to make things as simple Quantex Research Ltd, Manchester UK as possible but not simpler" [log in to unmask] [log in to unmask] +44(0)161 445 4951/G:+44(0)7717758675 www.quantex-research.com