Genetic Programming: 24th European Conference, EuroGP 2021, Held as Part of EvoStar 2021, Virtual Event, April 7–9, 2021, Proceedings (Theoretical Computer Science and General Issues)

This book constitutes the refereed proceedings of the 24th European Conference on Genetic Programming, EuroGP 2021, held as part of Evo*2021, as Virtual Event, in April 2021, co-located with the Evo*2021 events, EvoCOP, EvoMUSART, and EvoApplications. The 11 revised full papers and 6 short papers presented in this book were carefully reviewed and selected from 27 submissions. The wide range of topics in this volume reflects the current state of research in the field. The collection of papers cover interesting topics including developing new operators for variants of GP algorithms, as well as exploring GP applications to the optimisation of machine learning methods and the evolution of complex combinational logic circuits. Preface Organization Committee Contents Long Talks Quality Diversity Genetic Programming for Learning Decision Tree Ensembles 1 Introduction 2 Background 2.1 Decision Tree Diversity 2.2 Decision Tree Induction 2.3 Evolutionary Decision Tree Induction 2.4 QD Algorithms 2.5 Ensemble Prediction 3 Implementation 3.1 Structure Restrictions 3.2 Bagging 3.3 Random Trees 3.4 QD Trees 4 Experiments 5 Results 5.1 Algorithm Comparisons 5.2 Evolutionary Algorithm 5.3 Binary Diversity 5.4 Ensemble Size 6 Analysis 7 Conclusion References Progressive Insular Cooperative GP 1 Introduction 2 GP for Multiclass Classification 3 Progressive Insular Cooperative GP 3.1 Specialists Evolution Components 3.2 Teams Evolution Components 4 Experimental Study 4.1 Test Problems and Experimental Settings 4.2 Experimental Results 5 Conclusions and Future Work References Regenerating Soft Robots Through Neural Cellular Automata 1 Introduction 2 Related Work 2.1 Evolved Virtual Creatures 2.2 Cellular Automata 2.3 Shape Changing Robots and Damage Recovery 3 Growing Soft Robots with Neural Cellular Automata 3.1 Soft Robot Simulator 3.2 Genetic Algorithm 4 Results 4.1 Evolving 2D Soft Robots 4.2 Evolving 3D Soft Robots 4.3 Regenerating Soft Robots 5 Discussion and Future Work References Inclusive Genetic Programming 1 Introduction 2 Inclusive Evolutionary Process 2.1 Niches Creation 2.2 Inclusive Reproduction and Inclusive Tournament 3 Algorithms Setup and Results 3.1 Benchmarks 3.2 SGP and IGP Settings 3.3 Results 4 Conclusions References Towards Incorporating Human Knowledge in Fuzzy Pattern Tree Evolution 1 Introduction 2 Background 2.1 XAI 2.2 Grammatical Evolution 2.3 Fuzzy GE 3 Explainable GE 3.1 Reducing Size of Trees 3.2 FPT Representation 3.3 Human in the Loop 4 Experimental Setup 4.1 GE Parameters 4.2 Fitness Function 4.3 Fairness Benchmarks 4.4 Expert Validation 5 Results 6 Conclusion References Evolutionary Neural Architecture Search Supporting Approximate Multipliers 1 Introduction 2 Related Work 3 Evolutionary NAS with Approximate Circuits 3.1 CNN Representation 3.2 Genetic Operators 3.3 Fitness Functions 3.4 Search Algorithm 3.5 NAS with Approximate Multipliers 4 Results 4.1 The Role of Approximate Multipliers in NAS 4.2 Execution Time 4.3 Comparison with Other Similar Designs 5 Conclusions References Automatic Design of Deep Neural Networks Applied to Image Segmentation Problems 1 Introduction 2 Related Work 2.1 Image Segmentation and Edge Detection 2.2 U-Net 3 Dynamic Structured Grammatical Evolution 3.1 Grammar 3.2 DSGE Encoding 3.3 Mapping 3.4 Search Engine 4 Automatic Design of U-Nets 4.1 U-Net Mirror Grammar 4.2 Genetic Operators 4.3 Evaluation 5 Experiments 5.1 BSDS500 Dataset 5.2 ISBI12 Dataset 5.3 Evolutionary Optimization 5.4 Results 6 Conclusion References On the Influence of Grammars on Crossover in Grammatical Evolution 1 Introduction 2 Grammatical Evolution 2.1 Mapping 2.2 Crossover 2.3 Grammar Design 3 Experiments 3.1 Grammars 3.2 Static Analysis 3.3 Dynamic Analysis 3.4 Effects During Evolutionary Search 4 Conclusions References On the Generalizability of Programs Synthesized by Grammar-Guided Genetic Programming 1 Introduction 2 Related Work 2.1 GP-Based Program Synthesis 2.2 Selection Methods 3 Methodology 3.1 Software Metrics 3.2 Benchmark Problems 3.3 Grammar-Guided GP Approach 4 Experiments and Results 4.1 Comparison of Success Rates 4.2 Analysis with Software Metrics 4.3 Software Metric-Based Program Selection 5 Conclusions and Future Work References Evolution of Complex Combinational Logic Circuits Using Grammatical Evolution with SystemVerilog 1 Introduction 2 Background 2.1 Cartesian Genetic Programming and Grammatical Evolution 3 Related Work 3.1 Functional-Level Evolution 3.2 Problem Decomposition 3.3 Increasing Evolvability Through the Improvement of Genetic Operators 4 Proposed Approach and Experimental Design 4.1 Problems 4.2 Evolutionary Parameters 4.3 Training and Testing 5 Results and Discussion 6 Conclusion and Future Work References Evofficient: Reproducing a Cartesian Genetic Programming Method 1 Introduction 2 Background 3 Reproduction 3.1 Polishing-Up Schema 3.2 Fitness Evaluation Schema 3.3 Full-Stack Comparison 3.4 Random Search Baseline 4 Accelerating Neural Architecture Search Using Accuracy Prediction 4.1 Time Series Based Accuracy Prediction Accelerated CGP-CNN 5 Adaptability 5.1 Full-Dataset-Scenario 5.2 Small-Dataset-Scenario 6 Conclusion References Short Talks Software Anti-patterns Detection Under Uncertainty Using a Possibilistic Evolutionary Approach 1 Introduction 2 The Proposed Approach: ADIPOK 2.1 A Brief Overview of ADIPOK 2.2 Construction of PBEs 2.3 Chromosome Encoding 2.4 Fitness Function 2.5 Mating Selection Operator 2.6 Crossover and Mutation Operators 2.7 Optimized Possibilistic Detectors Application Module 3 Experimental Validation 3.1 Performance Metrics and Statistical Testing 3.2 Parameters Settings 4 Comparative Results 4.1 Results for RQ1 4.2 Results for RQ2 5 Related Work 6 Conclusions and Future Work References Probabilistic Grammatical Evolution 1 Introduction 2 Grammatical Evolution 2.1 Representation Variants 3 Probabilistic Grammatical Evolution 4 Experimental Analysis 4.1 Problem Description 4.2 Parameters 5 Results 6 Conclusion References Evolving Allocation Rules for Beam Search Heuristics in Assembly Line Balancing 1 Introduction 2 Station-Oriented Heuristics for the ALWABP 2.1 Iterated Probabilistic Beam Search 3 Genetic Programming Applied to Task Selection Rules 3.1 Crossover and Mutation Strategies 3.2 Strategies to Enhance the Search for Good Programs 4 Computational Experiments 4.1 Instances 4.2 Experiment 1: Calibration of the Genetic Algorithm 4.3 Experiment 2: Performance of Evolved Rules in Beam Searches 4.4 Experiment 3: Performance of the Best Evolved Rules 5 Conclusions References Incremental Evaluation in Genetic Programming 1 Background: Genetic Programming Evolving Functions 2 Incremental Evaluation: How Does It Work 3 Implementing Incremental Evaluation 4 Discussion: Population Convergence 4.1 Why Does Incremental Evaluation Work? 4.2 Implications for Software Engineering 5 Conclusion: Information Loss is Essential to Converge References Mining Feature Relationships in Data 1 Introduction 2 Background 2.1 Related Work 3 Proposed Method: GP-FRM 3.1 Overall Algorithm 3.2 Fitness Function 3.3 Preventing the Discovery of Naïve Relationships 4 Experiment Design 5 Results 5.1 Convergence Curves 6 Further Analysis 6.1 Analysis of Selected Features 6.2 Analysis of Evolved Relationships 7 Conclusion References Getting a Head Start on Program Synthesis with Genetic Programming 1 Introduction 2 Related Work 3 Method 3.1 Formalization 3.2 GP 3.3 Grammatical Evolution 4 Experiments 4.1 Solutions of Similar Problems 4.2 Experimental Setup 4.3 Experimental Design 5 Results 5.1 Discussion 6 Conclusions and Future Work References Author Index