Univ.-Prof. DI Dr. Alexander Felfernig Institute for Software Technology 1 Intelligent Techniques for Software Product Line Engineering 2 nd International Workshop on Formal Methods and Analysis in Software Product Line Engineering Univ.-Prof. DI Dr. Alexander Felfernig Institute for Software Technology Graz University of Technology, Austria alexander.felfernig@ist.tugraz.at
32
Embed
Intelligent Techniques for Software Product Line Engineering
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Univ.-Prof. DI Dr. Alexander Felfernig
Institute for Software Technology
1
Intelligent Techniques forSoftware Product Line Engineering
2nd International Workshop on Formal Methods and Analysis in Software Product Line Engineering
Univ.-Prof. DI Dr. Alexander Felfernig
Institute for Software TechnologyGraz University of Technology, Austria
Pre-informed Customers(knowledge about product/service)
CorporateMemory(standard levelof service)
IncreasedTrust(explanations)
Univ.-Prof. DI Dr. Alexander Felfernig
Institute for Software Technology
7
Motivation
2. Debuggingthe
Knowledge Base
1. Modeling the Knowledge
Base
3. Supporting the
User
Goal of this Presentation
Provide answers to the following questions:• How to develop constraint-based systems efficiently?• How to improve the quality of the underlying user interfaces?
Important Aspects:
Univ.-Prof. DI Dr. Alexander Felfernig
Institute for Software Technology
8
Modeling
2. DebuggingKnowledge Bases
1. ModelingKnowledge Bases
3. Supporting the User
Challenge: Knowledge representation languages have to be understandable (CSP, DCSP, GCSP)
[Mittal and Falkenhainer 1990], [Fleischanderl et al. 1998]
Approach: Domain-specific modeling concepts[Felfernig et al. 2000, 2003, 2006, 2007]
Motivation • Modeling
Result (among others): Integration to commercial environment
A. Felfernig. Standardized ConfigurationKnowledge Representations as Tech-nological Foundation for Mass Customi-zation, IEEE Transactions on Engi-neering Mgmt., 54(1), pp. 41-56, 2007.
Motivation • Modeling
Univ.-Prof. DI Dr. Alexander Felfernig
Institute for Software Technology
10
ConfigWorks Environment
A. Felfernig, G. Friedrich, D. Jannach, and M. Zanker.An Integrated Environment for the Development of Knowledge-based Recommender Applications, Intl. Journal of ElectronicCommerce (IJEC), 11(2):11-34, 2006.
Applications
Testing & Debugging
Product Properties
Questions to Customer
Dialog Modeling
Constraints
Motivation • Modeling
Univ.-Prof. DI Dr. Alexander Felfernig
Institute for Software Technology
11
Example: Generated Application
Motivation • Modeling
A. Felfernig and A. Kiener, Knowledge-basedInteractive Selling of Financial Services with FSAd-visor, 17th Innovative Applications of ArtificialIntelligence Conference (IAAI05). Pittsburgh, Penn-sylvania: AAAI Press, pp. 1475–1482, 2005.
Univ.-Prof. DI Dr. Alexander Felfernig
Institute for Software Technology
12
Debugging
2. DebuggingKnowledge Bases
1. Modeling Knowledge Bases
3. Supporting the User
Challenge: Complex knowledge bases with frequent maintenance steps [Barker et al. 1989], [Fleischanderl 2002], [Felfernig et al. 2007]
Approach: Automated Debugging of Knowledge Bases[Felfernig et al. 2004, 2006]
Motivation • Modeling • Debugging
Result: Prototypical integration to commercial environment
[Felfernig et al. 2006]
Univ.-Prof. DI Dr. Alexander Felfernig
Institute for Software Technology
13
Study @TUGraz: Are we able to identify faulty constraints in a knowledge base?
Further Study: A. Felfernig, Reducing Developmentand Maintenance Efforts for Web-based RecommenderApplications, International Journal of Web Engi-neering and Technology, 3(3), pp. 329-351, 2007.
Scenario: ti T = {t1, …, tt}: inconsistent (CKB ti) Conflict Set (CS): CS CKB s.t. ti T: inconsistent (CS ti)Minimal (CS): ¬CS‘ s.t. CS‘ CSExplanation (): CKB: (CKB - ) ti consistent ti TMinimal (): ¬‘ s.t. ‘
c1 c2 c3 … cq
Should Be: Test cases acceptedBut: some not accepted
1 2 d
t1 t2 t3 … tt
A. Felfernig, G. Friedrich, D. Jannach, and M. Stumpt-ner, Consistency-based Diagnosis of configuration know-ledge bases, AI Journal, 152(2):213–234, 2004.
…
Motivation • Modeling • Debugging
R. Reiter. A theory of diagnosisfrom first principles, AI Journal,23(1):57–95, 1987.
A. Felfernig, C. Zehentner, and P. Blazek. CoreDiag:Eliminating Redundancy in Constraint Sets, 22ndInternational Workshop on Principles of Diagnosis,Munich, Germany, to appear, 2011.
A constraint ci is called redundant iff CKB – {ci} ╞ ci. If this condition is not fulfilled, ci is said to be non-redundant.
U. Junker. QuickXplain: Preferredexplanations and relaxations for over-constrained problems. AAAI‘04, pp:167–172, San Jose, CA, 2004.
Motivation • Modeling • Debugging
Configuration Benchmarks Most knowledge bases are redundant!
Univ.-Prof. DI Dr. Alexander Felfernig
Institute for Software Technology
16
ConfigWorks Debugging Environment
Motivation • Modeling • Debugging
Debugging
Regression Testing FaultyConstraint
A. Felfernig, G. Friedrich, D. Jannach, and M. Zanker.An Integrated Environment for the Development ofKnowledge-based Recommender Applications, Intl. Journalof Electronic Commerce (IJEC), 11(2):11-34, 2006.
Univ.-Prof. DI Dr. Alexander Felfernig
Institute for Software Technology
17
User Support
2. DebuggingKnowledge Bases
1. ModelingKnowledge Bases
3. Supportingthe User
Challenge: Interaction support has potential for improvement,for example, „no solution could be found“
[Felfernig et al. 2004]
Approach: Explanations show the path to the solution[Felfernig et al. 2006, 2009, 2011]
Motivation • Modeling • Debugging • User Support
Result: Integration in commercial applicationsAlgorithmic improvements (FastDiag)
Should Be: Consistent RequirementsBut: Requirements inconsistent
1 2 d
cq+1 cq+2 cq+3 … cs
…
Motivation • Modeling • Debugging • User Support
CKB
CR
A. Felfernig, G. Friedrich, D. Jannach, and M. Stumpt-ner, Consistency-based Diagnosis of configuration know-ledge bases, AI Journal, 152(2):213–234, 2004.
R. Reiter. A theory of diagnosisfrom first principles, AI Journal,23(1):57–95, 1987.
• Explanations with high Probability [DeKleer 1990]
• Utility-based Explanations [Jannach and Liegl 2006]
• Corrective Explanations [O‘Sullivan et al. 2005]
• Representative Explanations [O‘Sullivan et al. 2007]
• kNN-based Explanations [Felfernig et al. 2009]
• Ensemble-based Explanations [Felfernig et al. 2011]
A. Felfernig, M. Schubert, and C. Zehentner. An Efficient Diagnosis Algorithm forInconsistent Constraint Sets, AI for Engineering Design, Analysis, and Manufacturing(AIEDAM), Cambridge University Press, 25(2):175—184, 2011.
► Existing Approaches: good prediction quality but inefficient.► FastDiag: Efficient determination of preferred explanations.
Recommender Systems (RS)Intelligent Navigation Support ML + RSIntelligent Maintenance Support ML + RS
User 4 would receive a recommendation for f4 (since the majority of nearest neighbors looked at f4).
Motivation • Modeling • Debugging • User Support • Research Issues
A. Falkner, A. Felfernig, and A. Haag. Recommendation Technologies for ConfigurableProducts, AI Magazine, AAAI, to appear, Fall 2011.
Univ.-Prof. DI Dr. Alexander Felfernig
Institute for Software Technology
25
Summary• Constraint-based systems are very popular (e.g., CRM, Online
Selling, Call Center, Project Management, Software Engineering)
• Challenge: large & complex knowledge bases
• Basis for efficient development processes:• domain-specific modeling languages• automated (personalized) debugging techniques
• Efficient user interaction on the basis of preferred explanations
• Future research focus: human-centered interaction with complex knowledge spaces.
Motivation • Modeling • Debugging • User Support • Summary
Univ.-Prof. DI Dr. Alexander Felfernig
Institute for Software Technology
26
Questions?
Univ.-Prof. DI Dr. Alexander Felfernig
Institute for Software Technology
27
References (1)A. Felfernig, G. Friedrich, D. Jannach, and M. Zanker, An Integrated
Environment for the Development of Knowledge-based Recommender Applications, International Journal of Electronic Commerce (IJEC), 11(2):11-34, 2006.
S. Russell and P. Norvig. Artificial Intelligence: A Modern Approach (2nd ed.), Prentice Hall, 2003.
E. Freuder. In the pursuit of the holy grail, Constraints: An International Journal, 2:57–61, 1997.
D. Benavides, S. Segura, A. Cortés: Automated analysis of feature models 20 years later: A literature review. Inf. Syst. 35(6): 615-636, 2010.
A. Felfernig and A. Kiener, Knowledge-based Interactive Selling of Financial Services with FSAdvisor, 17th Innovative Applications of Artificial Intelligence Conference (IAAI05). Pittsburgh, Pennsylvania: AAAI Press, 2005, pp. 1475–1482.
Univ.-Prof. DI Dr. Alexander Felfernig
Institute for Software Technology
28
References (2)A. Felfernig, G. Friedrich, D. Jannach, and M. Zanker, An Integrated
Environment for the Development of Knowledge-based Recommender Applications, International Journal of Electronic Commerce (IJEC), 11(2):11-34, 2006.
A. Felfernig. Standardized Configuration Knowledge Representations as Tech-nological Foundation for Mass Customization, IEEE Transactions on Engineering Mgmt., 54(1), pp. 41-56, 2007.
S. Mittal and B. Falkenhainer, Dynamic Constraint Satisfaction Problems, in National Conference on Artificial Intelligence (AAAI 90), Boston, MA, pp. 25–32, 1990.
A. Felfernig, G. Friedrich, and D. Jannach, UML ad domain specific language for the construction of knowledge-based configuration systems, International Journal on Software Engineering and Knowledge Engineering, 10(4):449–470, 2000.
A. Felfernig, G. Friedrich, D. Jannach, M. Stumptner, and M. Zanker, Configuration Knowledge Representations for Semantic Web Applications, Artificial Intelligence in Engineering, Design, Analysis and Manufacturing (AIEDAM), 17(2):31–50, 2003.
Univ.-Prof. DI Dr. Alexander Felfernig
Institute for Software Technology
29
References (3)G. Fleischanderl, G. Friedrich, A. Haselboeck, H. Schreiner, and M.
Stumptner, Configuring Large Systems Using Generative Constraint Satisfaction, IEEE Intelligent Systems, 13(4):59–68, 1998.
G. Fleischanderl. Suggestions from the Software Engineering Practice for Applying Consistency-Based Diagnosis to Configuration Knowledge Bases, 13th International Workshop on Principles of Diagnosis (DX-02), pp. 1—3, 2002.
V. Barker, D. O’Connor, J. Bachant, and E. Soloway, Expert systems for configuration at Digital: XCON and beyond, Communications of the ACM, vol. 32, no. 3, pp. 298–318, 1989.
A. Felfernig, G. Friedrich, D. Jannach, and M. Stumptner, Consistency-based Diagnosis of configuration knowledge bases, in Artificial Intelligence, 152(2), 2004, pp. 213–234.
A. Felfernig, Reducing Development and Maintenance Efforts for Web-based Recommender Applications, International Journal of Web Engi-neering and Technology, 3(3), pp. 329-351, 2007.
Univ.-Prof. DI Dr. Alexander Felfernig
Institute for Software Technology
30
References (4)R. Reiter. A theory of diagnosis from first principles, AI Journal, 23(1):57–
95, 1987.A. Felfernig, G. Friedrich, M. Schubert, M. Mandl, M. Mairitsch, and E.
Teppan. Plausible Repairs for Inconsistent Requirements, Proceedings of the 21st International Joint Conference on Artificial Intelligence (IJCAI'09), Pasadena, California, USA, pp. 791-796, 2009.
A. Felfernig, M. Schubert, and C. Zehentner. An Efficient Diagnosis Algorithm for Inconsistent Constraint Sets, AI for Engineering Design, Analysis, and Manufacturing (AIEDAM), Cambridge University Press, 25(2):175—184, 2011.
U. Junker. QuickXplain: Preferred explanations and relaxations for over-constrained problems. AAAI‘04, pp. 167–172, San Jose, CA, 2004.
J. de Kleer, A. Mackworth, and R.Reiter. Characterizing diagnoses and systems, AI Journal, 56(2-3):197-222, 1992.
Univ.-Prof. DI Dr. Alexander Felfernig
Institute for Software Technology
31
References (5)B. O’Callaghan, B. O’Sullivan, and E. C. Freuder. Generating Corrective
Explanations for Interactive Constraint Satisfaction. In CP’2005, pp. 445–459, 2005.
B. O’Sullivan, A. Papdopoulos, B. Faltings, and P. Pu. Representative Explanations for Over-Constrained Problems. AAAI’07, pp. 323–328, 2007.
D. Jannach. Finding Preferred Query Relaxations in Content-based Recommenders, IEEE Intelligent Systems Conf. (IS’2006), pp. 355-360, 2006.
A. Felfernig and M. Schubert. Personalizing Diagnoses for Inconsistent Constraint Sets, 22nd International Workshop on Principles of Diagnosis, Munich, Germany, to appear, 2011.
A. Falkner, A. Felfernig, and A. Haag. Recommendation Technologies for Configurable Products, AI Magazine, AAAI, to appear, Fall 2011.