Machine Learning for Dynamic Software Analysis: Potentials and Limits

1. Front Matter

   Pages I-IX

   PDF

2. Introduction

   1. Front Matter

      Pages 1-1

      PDF

   2. Machine Learning for Software Analysis: Models, Methods, and Applications

      • Amel Bennaceur, Karl Meinke

      Pages 3-49

3. Testing and Learning

   1. Front Matter

      Pages 51-51

      PDF

   2. Learning-Based Testing: Recent Progress and Future Prospects

      • Karl Meinke

      Pages 53-73

   3. Model Learning and Model-Based Testing

      • Bernhard K. Aichernig, Wojciech Mostowski, Mohammad Reza Mousavi, Martin Tappler, Masoumeh Taromirad

      Pages 74-100

   4. Testing Functional Black-Box Programs Without a Specification

      • Neil Walkinshaw

      Pages 101-120

4. Extensions of Automata Learning

   1. Front Matter

      Pages 121-121

      PDF

   2. Active Automata Learning in Practice

      • Falk Howar, Bernhard Steffen

      Pages 123-148

   3. Extending Automata Learning to Extended Finite State Machines

      • Sofia Cassel, Falk Howar, Bengt Jonsson, Bernhard Steffen

      Pages 149-177

   4. Inferring FSM Models of Systems Without Reset

      • Roland Groz, Adenilso Simao, Alexandre Petrenko, Catherine Oriat

      Pages 178-201

5. Integrative Approaches

   1. Front Matter

      Pages 203-203

      PDF

   2. Constraint-Based Behavioral Consistency of Evolving Software Systems

      • Reiner Hähnle, Bernhard Steffen

      Pages 205-218

   3. Logic-Based Learning: Theory and Application

      • Dalal Alrajeh, Alessandra Russo

      Pages 219-256

6. Back Matter

   Pages 257-257

   PDF

Machine learning of software artefacts is an emerging area of interaction between the machine learning and software analysis communities.  Increased productivity in software engineering relies on the creation of new adaptive, scalable tools that can analyse large and continuously changing software systems.  These require new software analysis techniques based on machine learning, such as learning-based software testing, invariant generation or code synthesis. Machine learning is a powerful paradigm that provides novel approaches to automating the generation of models and other essential software artifacts.  This volume originates from a Dagstuhl Seminar entitled "Machine Learning for Dynamic Software Analysis: Potentials and Limits” held in April 2016. The seminar focused on fostering a spirit of collaboration in order to share insights and to expand and strengthen the cross-fertilisation between the machine learning and software analysis communities.  The book provides an overview of the machine learning techniques that can be used for software analysis and presents example applications of their use. Besides an introductory chapter, the book is structured into three parts: testing and learning, extension of automata learning, and integrative approaches.

• Active learning
• Artificial intelligence
• Automated static analysis
• Computing methodologies
• Dynamic analysis
• Formal languages and automata theory
• Formal methods
• Machine learning
• Model development and analysis
• Semantics
• Software design
• Software engineering
• Specifications
• Theory and algorithms for application domains
• Theory of computation

• The Open University, Milton Keynes, United Kingdom

  Amel Bennaceur

• Technische Universität Darmstadt, Darmstadt, Germany

  Reiner Hähnle

• KTH Royal Institute of Technology, Stockholm, Sweden

  Karl Meinke

Policies and ethics