Handbook on Artificial Intelligence-Empowered Applied Software Engineering: Volume 1: Novel Methodologies to Engineering Smart Software Systems

Description

This book provides a structured overview of artificial intelligence-empowered applied software engineering. Evolving technological advancements in big data, smartphone and mobile software applications, the Internet of Things and a vast range of application areas in all sorts of human activities and professions lead current research towards the efficient incorporation of artificial intelligence enhancements into software and the empowerment of software with artificial intelligence. This book at hand, devoted to Novel Methodologies to Engineering Smart Software Systems Novel Methodologies to Engineering Smart Software Systems, constitutes the first volume of a two-volume Handbook on Artificial Intelligence-empowered Applied Software Engineering. Topics include very significant advances in (i) Artificial Intelligence-Assisted Software Development and (ii) Software Engineering Tools to develop Artificial Intelligence Applications, as well as a detailed Survey of Recent Relevant Literature.  Professors, researchers, scientists, engineers and students in artificial intelligence, software engineering and computer science-related disciplines are expected to benefit from it, along with interested readers from other disciplines. Foreword Preface Contents 1 Introduction to Handbook on Artificial Intelligence-Empowered Applied Software Engineering—VOL.1: Novel Methodologies to Engineering Smart Software Systems 1.1 Editorial Note 1.2 Book Summary and Future Volumes Bibliography for Further Reading Part I Survey of Recent Relevant Literature 2 Synergies Between Artificial Intelligence and Software Engineering: Evolution and Trends 2.1 Introduction 2.2 Methodology 2.3 The Evolution of AI in Software Engineering 2.4 Top Authors and Topics 2.5 Trends in AI Applications to Software Engineering 2.5.1 Machine Learning and Data Mining 2.5.2 Knowledge Representation and Reasoning 2.5.3 Search and Optimisation 2.5.4 Communication and Perception 2.5.5 Cross-Disciplinary Topics 2.6 AI-Based Tools 2.7 Conclusion References Part II Artificial Intelligence-Assisted Software Development 3 Towards Software Co-Engineering by AI and Developers 3.1 Introduction 3.2 Software Development Support and Automation Level by Machine Learning 3.2.1 Project Planning: Team Composition 3.2.2 Requirements Engineering: Data-Driven Persona 3.2.3 Design: Detection of Design Patterns 3.2.4 Categorization of Initiative and Level of Automation 3.3 Quality of AI Application Systems and Software 3.3.1 Metamorphic Testing 3.3.2 Improving Explainability 3.3.3 Systems and Software Architecture 3.3.4 Integration of Goals, Strategies, and Data 3.4 Towards Software Co-Engineering by AI and Developers 3.5 Conclusion References 4 Generalizing Software Defect Estimation Using Size and Two Interaction Variables 4.1 Introduction 4.2 Background 4.3 A Proposed Approach 4.3.1 Selection of Sample Projects 4.3.2 Data Collection 4.3.3 The Scope and Decision to Go with ‘Interaction’ Variables 4.3.4 Data Analysis and Results Discussion 4.3.5 The Turning Point 4.3.6 Models Performance—Outside Sample 4.4 Conclusion and Limitations 4.5 Future Research Directions 4.6 Annexure—Model Work/Details References 5 Building of an Application Reviews Classifier by BERT and Its Evaluation 5.1 Background 5.2 The Process of Building a Machine Learning Model 5.3 Dataset 5.4 Preprocessing 5.5 Feature Engineering 5.5.1 Bag of Words (BoW) [4] 5.5.2 FastText [5, 6] 5.5.3 Bidirectional Encoder Representations from Transformers (BERT) Embedding [7] 5.6 Machine-Learning Algorithms 5.6.1 Naive Bayes 5.6.2 Logistic Regression 5.6.3 BERT 5.7 Training and Evaluation Methods 5.8 Results 5.9 Discussion 5.9.1 Comparison of Classifier Performances 5.9.2 Performance of the Naive Bayes Classifiers 5.9.3 Performance of the Logistic Regression Classifiers 5.9.4 Visualization of Classifier Attention Using the BERT 5.10 Threats to Validity 5.10.1 Labeling Dataset 5.10.2 Parameter Tuning 5.11 Summary References 6 Harmony Search-Enhanced Software Architecture Reconstruction 6.1 Introduction 6.2 Related Work 6.3 HS Enhanced SAR 6.3.1 SAR Problem 6.3.2 HS Algorithm 6.3.3 Proposed Approach 6.4 Experimentation 6.4.1 Test Problems 6.4.2 Competitor approaches 6.5 Results and Discussion 6.6 Conclusion and Future Work References 7 Enterprise Architecture-Based Project Model for AI Service System Development 7.1 Introduction 7.2 Related Work 7.3 AI Servie System and Enterprise Architecture 7.3.1 AI Service System 7.3.2 Enterprise Architecture and AI Service System 7.4 Modeling Business IT Alignment for AI Service System 7.4.1 Generic Business–AI Alignment Model 7.4.2 Comparison with Project Canvas Model 7.5 Business Analysis Method for Constructing Domain Specific Business–AI Alignment Model 7.5.1 Business Analysis Tables 7.5.2 Model Construction Method 7.6 Practice 7.6.1 Subject Project 7.6.2 Result 7.7 Discussion 7.8 Conclusion References Part III Software Engineering Tools to Develop Artificial Intelligence Applications 8 Requirements Engineering Processes for Multi-agent Systems 8.1 Introduction 8.2 Background 8.2.1 Agents, Multiagent Systems, and the BDI Model 8.2.2 Requirements Engineering 8.3 Techniques and Process of Requirements Engineering for Multiagent Systems 8.3.1 Elicitation Requirements Techniques for Multiagent Systems 8.3.2 Requirements Engineering Processes for Multiagent Systems 8.3.3 Requirements Validation 8.4 Conclusion References 9 Specific UML-Derived Languages for Modeling Multi-agent Systems 9.1 Introduction 9.2 Backgroud 9.2.1 UML 9.2.2 Agents, Multiagent Systems, and the BDI Model 9.2.3 BDI Models 9.3 AUML—Agent UML 9.4 AORML—Agent-Object-Relationship Modeling Language 9.4.1 Considerations About AORML 9.5 AML—Agent Modeling Language 9.5.1 Considerations About AML 9.6 MAS-ML—Multiagent System Modeling Language 9.6.1 Considerations About MAS-ML 9.7 SEA-ML—Semantic Web Enabled Agent Modeling Language 9.7.1 Considerations 9.8 MASRML—A Domain-Specific Modeling Language for Multi-agent Systems Requirements 9.8.1 Considerations References 10 Methods for Ensuring the Overall Safety of Machine Learning Systems 10.1 Introduction 10.2 Related Work 10.2.1 Safety of Machine Learning Systems 10.2.2 Conventional Safety Model 10.2.3 STAMP and Its Related Methods 10.2.4 Standards for Software Lifecycle Processes and System Lifecycle Processes 10.2.5 Social Technology Systems and Software Engineering 10.2.6 Software Layer Architecture 10.2.7 Assurance Case 10.2.8 Autonomous Driving 10.3 Safety Issues in Machine Learning Systems 10.3.1 Eleven Reasons Why We Cannot Release Autonomous Driving Cars 10.3.2 Elicitation Method 10.3.3 Eleven Problems on Safety Assessment for Autonomous Driving Car Products 10.3.4 Validity to Threats 10.3.5 Safety Issues of Automatic Operation 10.3.6 Task Classification 10.3.7 Unclear Assurance Scope 10.3.8 Safety Assurance of the Entire System 10.3.9 Machine Learning and Systems 10.4 STAMP S&S Method 10.4.1 Significance of Layered Modeling of Complex Systems 10.4.2 STAMP S&S and Five Layers 10.4.3 Scenario 10.4.4 Specification and Standard 10.5 CC-Case 10.5.1 Definition of CC-Case 10.5.2 Technical Elements of CC-Case 10.6 Measures for Autonomous Driving 10.6.1 Relationship Between Issues and Measures Shown in This Section 10.6.2 Measure 1: Analyze Various Quality Attributes in Control Action Units 10.6.3 Measure 2: Modeling the Entire System 10.6.4 Measure 3: Scenario Analysis and Specification 10.6.5 Measure 4: Socio-Technical System 10.7 Considerations in Level 3 Autonomous Driving 10.7.1 Example of Autonomous Driving with the 5-layered Model of STAMP S&S 10.8 Conclusion References 11 MEAU: A Method for the Evaluation of the Artificial Unintelligence 11.1 Introduction 11.2 Machine Learning and Online Unintelligence: Improvisation or Programming? 11.3 The New Paradigm of Information from Digital Media and Social Networks 11.4 Numbers, Images and Texts: Sources of Errors, Misinformation and Unintelligence 11.5 MEAU: A Method for the Evaluation of the Artificial Unintelligence 11.6 Results 11.7 Lessons Learned 11.8 Conclusions Appendix 1 Appendix 2 Appendix 3 Appendix 4 References 12 Quantum Computing Meets Artificial Intelligence: Innovations and Challenges 12.1 Introduction 12.1.1 Benefits of Quantum Computing for AI 12.2 Quantum Computing Motivations 12.2.1 What Does ``Quantum'' Mean? 12.2.2 The Wave-Particle Duality 12.2.3 Qubit Definition 12.2.4 The Schrödinger Equation 12.2.5 Superposition 12.2.6 Interference 12.2.7 Entanglement 12.2.8 Gate-Based Quantum Computing 12.3 Quantum Machine Learning 12.3.1 Variational Quantum Algorithms 12.3.2 Data Encoding 12.3.3 Quantum Neural Networks 12.3.4 Quantum Support Vector Machine 12.3.5 Variational Quantum Generator 12.4 Quantum Computing Limitations and Challenges 12.4.1 Scalability and Connectivity 12.4.2 Decoherence 12.4.3 Error Correction 12.4.4 Qubit Control 12.5 Quantum AI Software Engineering 12.5.1 Hybrid Quantum-Classical Frameworks 12.5.2 Friction-Less Development Environment 12.5.3 Quantum AI Software Life Cycle 12.6 A new Problem Solving Approach 12.6.1 Use Case 1: Automation and Transportation Sector 12.6.2 Use Case 2: Food for the Future World 12.6.3 Use Case 3: Cheaper Reliable Batteries 12.6.4 Use Case 4: Cleaner Air to Breathe 12.6.5 Use Case 5: AI-Driven Financial Solutions 12.7 Summary and Conclusion References