Transcript Slide 1
The 23rd International Conference on Software Engineering and Knowledge Engineering (SEKE 2011) A Virtual Catalyst in the Knowledge Acquisition Process Geraldo Boz Junior, Tecpar Milton Pires Ramos, Tecpar Gilson Yukio Sato, UTFPR Julio Cesar Nievola, PUCPR Emerson Cabrera Paraiso, PUCPR Introduction Virtual Catalyst Noctua Tool Features Structure Introduction • AI – Artificial Intelligence • Knowledge Based Systems • Analysis, diagnosis • Preservation of knowledge • KA – Knowledge Acquisition • Expert + Knowledge Engineer • Problems: deadlines, expenses, time availability, knowledge representation • CKC – Collaborative Knowledge Construction • Multiple remote collaborators • Problems: authorship, validation • Noctua = AI + KA + CKC 3 Knowledge Acquisition (KA) Definition • Conceptual Knowledge x Procedural Knowledge [1] • Knowledge Acquisition is the explanation and the capture of knowledge in a structured format. [2] KA techniques • Interviews, simulation, scenarios [1] [3] Knowledge Representation • Knowledge Pages, Production Rules [1] [2] Problems • Faulty documentation, elicitation difficulty, disorganization, ignorance, availability [4] 4 Collaborative Knowledge Construction (CKC) Distance Collaboration • Synchronous collaborative sessions x asynchronous collaboration [5] Incentive to Collaboration • Productivity awards, reputation inside the group, social translucence [6] Effectiveness: stimulus and measurements • Fostering interaction, distribution of roles, metacognition [7] • Quantity of logins, produced artifacts, quantity of messages and comments, etc. [8] Tool characteristics • Web, simple, forum, questions, synchronous/asynchronous, authorship, search [9] Building consensus • Authority, consensus [10] 5 Proactivity (CKC) Intelligent Systems • Ability to understand and act on the environment according to their own objectives. [5] • Perceptions and actions; memory, knowledge and goals; planning and decision making [11] Profile of collaborators • Interaction vary with interests, knowledge, history of activities [6] Artificial element action • Familiarization, discussion [8] • Types of questions (“Evaluate...” , “What if...?”) [12] 6 Introdution Virtual Catalyst Noctua Tool Features Structure Noctua Project Hyper Hiperglossário Glossary Base Rule deBase Regras Input Variables Internal Variables Output Conclusions Auxiliary Conclusions Constants Terminal Conclusions Inference Engine 8 Expert(s) Rule Base Hyper Glossary Instant Messages ? Profiles Questions Images ! Log Knowledge Engineer(s) Comments Tags Project Memory 9 Introduction Virtual Catalyst Noctua Tool Features Structure Knowledge Page 11 Production Rule 12 Introduction Virtual Catalyst Noctua Tool Features Structure Tool Features Expert(s) Rule Base Hyper Glossary Instant Messages Questions Images Log Knowledge Engineer(s) ? Profiles Comments ! Tags Project Memory Virtual Catalyst 14 Catalyst Action Rule X If Profile Tags SCC >= SCC_normal_limit SCC < SCC_high_limit Then high SCC Rule Y Rules If DIM >= lactation_initial_phase_limit DIM < lactation_end_limit Then last phase lactation 15 Catalyst Action Rule ? If SCC >= SCC_normal_limit SCC < SCC_high_limit DIM >= lactation_initial_phase_limit Then ??? Mr. Expert, is it possible to conclude something from these conditions? 16 Introduction Virtual Catalyst Noctua Tool Features Structure Experimentation Project Gourmet Entries/Rules http://projetos.dia.tecpar.br/noctua instigated 17% Pairing Foods and Wine • 9 distant collaborators • 111 questions made by Noctua • 204 instant messages • 50 rules and entries (17% instigated) • 60 opinions validating knowledge (23% instigated) spontaneous 83% Opinions instigated This experimentation inspired an improvement in the 23% tool, which now also integrates input variables to the spontaneous rules and entries of the project. 77% 18 Expected Results • A method for Knowledge Acquisition with characteristics of Collaborative Knowledge Construction and a Virtual Catalyst. • More efficient Knowledge Acquisition • Decrease the need to face meetings • Lower costs • Shorter development time • Procedural knowledge integrated with conceptual knowledge 19 Conclusion Work already done • Theoretical foundation • Defining tool features • Development of the tool (Noctua) Work in progress • More experiments • Assessment of experiment results • New tool features 20 References [1] MILTON, N.R Knowledge Acquisition in Practice, Springer-Verlag London Limited, 2007 [2] ROLSTON, D.W. Principles of Artificial Intelligence and Expert Systems Development. McGraw-Hill Book Co, 1988. [3] GROVER, M.D. A Pragmatic Knowledge Acquisition Methodology. Psychological Review, 1982, pp. 1-3. [4] MASTELLA, L.S. Técnicas de Aquisição de Conhecimento para Sistemas Baseados em Conhecimento, UFRGS, 2004. [5] SCHWARTZ, D.G. Encyclopedia of Knowledge Management. Idea Group Reference, 2006. [6] NABETH, T.; RODA, C.; ANGEHRN, A. e MITTAL, P. Using artificial agents to stimulate participation in virtual communities. ADIS International Conference CELDA (Cognition and Exploratory Learning in Digital Age), 2005, pp. 2-5. [7] PETTENATI, M.C. e RANIERI, M. Informal learning theories and tools to support knowledge management in distributed CoPs. Proceedings of the 1st International Workshop on Building Technology Enhanced Learning solutions for Communities of Practice, held in conjunction with the 1st European Conference on Technology Enhanced Learning Crete, Greece, 2006, pp. 345-355. [8] ANGEHRN, A.A. Designing Intelligent Agents for Virtual Communities. CALT Report 11-2004, 2004, pp. 1-29. [9] NOY, N.F.; CHUGH, A. e ALANI, H. The CKC Challenge: Exploring Tools for Collaborative Knowledge Construction. IEEE Intelligent Systems, vol. 23, 2008, pp. 64-68. [10] DIENG, R.; CORBY, O.; GIBOIN, A.; GOLEBIOWSKA J.; MATTA N. e RIBIÈRE M. Méthodes et outils pour la gestion des connaissances. Dunod, 2000. [11] KENDAL, S e CREEN, M. An Introduction to Knowledge Engineering. Springer-Verlag London Limited, 2007. [12] MCGRAW, K. e HARBISON-BRIGGS K. Knowledge acquisition: principles and guidelines. Prentice-Hall, Inc. Upper Saddle River, NJ, USA, 1989. 21 The 23rd International Conference on Software Engineering and Knowledge Engineering (SEKE 2011) A Virtual Catalyst in the Knowledge Acquisition Process Thank you! Geraldo Boz Junior, Tecpar Milton Pires Ramos, Tecpar Gilson Yukio Sato, UTFPR Julio Cesar Nievola, PUCPR Emerson Cabrera Paraiso, PUCPR