140 likes | 158 Views
1 4 th Workshop on “Software Engineering Education and Reverse Engineering” Sinaia, Romania August 2 4 th – 3 0th 201 4. GoF design patterns applied to artificial intelligence methods Assist. Prof. Dr. Dušanka Bošković University of Sarajevo. Content.
E N D
14th Workshop on “Software Engineering Education and Reverse Engineering” Sinaia, Romania August 24th– 30th 2014 GoF design patterns applied to artificial intelligence methods Assist. Prof. Dr. Dušanka BoškovićUniversity of Sarajevo
Content • Background: Integrating Software Engineering and Artificial Intelligence • DP implementation cases • Conclusion
Background • Benefits: • Fuzzy inferencing is suitable for solving real world problems: operating with fuzzy and non-complete data. • Adaptive systems development needs application of machine learning techniques. • Problems: • AI research focused to solving specific and particular problems, • SE distrust black-box approach and complexity of AI methods • Way to go: • AI: need to elevate abstraction level of solutions from a particular problem to a strategy that solves a class of problems • SE: It is not enough to embed AI method; integration of AI methods in SE processes and techniques
Integrating AI and SE • Building generic models for AI • Agents and Multi-agent systems (MAS) • Reasoning - Fuzzy inferencing system (FIS) • Learning - Artificial neural networks (ANN)
FIS model • Class diagram
FIS model • Fuzzy set generalization
ANN model • Class diagram • Feedforward NN with Backpropagation
Design patterns • Factory
Design patterns • Factory
Design patterns • Strategy
Design patterns • Composition
JAVA - Design patterns • Iterator • Observer
Conclusion • Bridging AI and SE • Making AI solutions understandible to SE • Introducing new examples for students
Thank you for your attention • Questions?