Received: from GLINDA.OZ.CS.CMU.EDU by A.GP.CS.CMU.EDU id aa13835; 17 Jul 95 14:43:45 EDT Date: Mon, 17 Jul 95 14:42:09 EDT From: AI.Repository@GLINDA.OZ.CS.CMU.EDU To: ai+ai-postdoc@cs.cmu.edu Subject: Postdoc: Fuzzy AI, Knowledge Engineering at Univ. of Bristol Sender: ai@A.GP.CS.CMU.EDU From: entpm@zeus.bris.ac.uk (TP. Martin) Subject: Research Opportunity (University of Bristol) Organization: University of Bristol, England Date: Fri, 7 Jul 1995 15:18:25 GMT Research Opportunities in Fuzzy Artificial Intelligence The Artificial Intelligence group at the University of Bristol (headed by Professor J.F. Baldwin) has a vacancy for a postdoctoral research assistant and a postgraduate research student to work on the topics described below. 1. PhD Case Studentship in Knowledge Engineering A student with a good honours degree in Mathematics, Physics, Computer Science or Engineering is required to work on the development of a fuzzy intelligent data browser for obtaining knowledge from large volume data. This work will be done in co-operation with DRA, Mavern. Starting date is October 1995. We live in the age of information - large amounts of data for a wide range of subject matter is there for the collecting - but data is numbers and intelligence is needed to make sense of it and convert it into worthwhile knowledge. Scientists, engineers, managers, business decision makers from many disciplines collect large amounts of data to be used to determine models for their applications. The data can represent pictures and graphics, sound, numbers and text. To know something is to understand the relationship between parts or features or concepts, to relate to a context, to obtain models, to use to infer. To discover is to find these relationships, to hypothesise and to justify. The models may be in mathematical form as equations, in logic form as propositions, in linguistic form as natural language statements, probabilistic in nature and use fuzzy representations. These models provide a summarisation of the data - a high level form of data compression. A data browser can answer queries with reference to a large database. The database can contain uncertain, vague and fuzzy attribute values. To answer a query some form of interpolation between near matching appropriate entries in the database may be required. This interpolation can take the form of forming fuzzy, probabilistic, mixed fuzzy probabilistic rules from the relevant data base entries. The browser effectively and intelligently fills in for missing information, constructs partial matchings and uses these to answer the query using a form of case based reasoning. The browser will be useful for many applications such as data mining, pattern recognition, fuzzy control, databases containing uncertainties etc. and financial modelling. 2. A Post-Doctoral Research Assistant is required to work on an EPSRC funded project "Mathematical Modelling and Knowledge Engineering for Engineering Design" Candidates should have experience with some of the following: - knowledge based systems - logic programming - mathematical modelling - uncertainty in AI - Mathematica - user friendly interfaces. This project is concerned with the development of a software environment for intelligent engineering analysis and design. The integration of mathematical models with logic-style knowledge bases should provide an intelligent working environment for the Engineer to use as a software laboratory in conjunction with the hardware laboratory for design purposes. This is not a simple concatenation of mathematics software with artificial intelligence software but represents an integration in which the mathematics aids the intelligence of the AI software and the artificial intelligence aids the validity of the mathematics software. The Engineer uses mathematical models, CAD software, intuition based on experience, simplifications and approximations, assumptions and design creativity when designing a new product. At present the Engineer uses NAG routines to perform numerical calculations for evaluating the mathematical model, graphics packages to see the results and specialised CAD software to help with the design. The Engineer is responsible for determining which models and assumptions are applicable to answer a given question, for providing explanations of why a certain result occurred, for setting up new conditions, assumptions, models and links with software packages to answer "what if" questions. This is in addition to the creativity expected in constructing the necessary changes in design for the next iteration of the iterative design process. The Engineer is expected to memorise relevant information or know how to find it. Two software systems exist today which the Engineer should learn to exploit far more intelligently than has been done to date. These are the mathematical symbol manipulating program Mathematica and the Artificial Intelligence language FRIL. Each by itself provides a powerful computing tool for the engineer but each one also has missing elements. Taken together, with one being able to ask questions of the other and data able to pass freely from one to the other, the combination will provide a powerful intelligent design environment. Supervision The research will be supervised by Prof J.F. Baldwin, Professor of Artificial Intelligence, EPSRC Senior Research F ellow and Dr. T.P. Martin Further Information: http://www.fen.bris.ac.uk/engmaths/research/aigroup/resopps.html or contact Jim.Baldwin@bristol.ac.uk Trevor.Martin@bristol.ac.uk _______________________________________________________________________________ Advanced Computing Research Centre || Phone: +44 117 9288200 Department of Engineering Maths || Fax: +44 117 9251154 University of Bristol, BS8 1TR, UK || email: trevor.martin@bris.ac.uk_______________________________________________________________________________ ------------------------------------------------------------------------------- This message | Submissions ai+ai-postdoc@cs.cmu.edu was sent via | Subscribe/Unsubscribe ai+query@cs.cmu.edu the AI-POSTDOC | Available mailing lists include mailing list. | AI-JOBS, LISP-JOBS, PROLOG-JOBS, AI-POSTDOC, AI-PREDOC