AIxIA 2022 – Advances in Artificial Intelligence XXIst International Conference of the Italian Association for Artificial Intelligence. AIxIA 2022, Udine, Italy, November 28 – December 2, 2022 Proceedings

Description

Preface Organization Contents Hybrid Approaches The PSyKE Technology for Trustworthy Artificial Intelligence 1 Introduction 2 The PSyKE framework 2.1 Functionalities and Main Components 2.2 Architecture and API 3 Examples 3.1 Classification: The Iris Data Set 3.2 Regression: The Combined Cycle Power Plant Data Set 3.3 PSyKE GUI 4 Impact 5 Conclusion References A Declarative Approach to Contrast Pattern Mining 1 Introduction 2 Preliminaries 2.1 Contrast Pattern Mining in Brief 2.2 Answer Set Programming in a Nutshell 3 Mining Contrast Patterns with ASP 4 An Application in Digital Forensics 4.1 The DigForASP Dataset 4.2 Preprocessing and ASP Encoding of the Dataset 4.3 Experiments 5 Final Remarks References Graphs and Networks Approximate Inference in Probabilistic Answer Set Programming for Statistical Probabilities 1 Introduction 2 Related Work 3 Background 4 Approximate Inference for PASTA Programs 5 Experiments 6 Conclusions References Decision Trees with a Modal Flavor 1 Introduction 2 Preliminaries 3 Modal Decision Trees 4 Applications 5 Conclusions References Assisted Process Knowledge Graph Building Using Pre-trained Language Models 1 Introduction 2 Related Work 3 Use Case 4 Knowledge Extraction from Text via In-Context Learning 4.1 In-Context Learning 4.2 Implementing the Approach 5 Empirical Assessment 5.1 The Tasks 5.2 Experimental Setting 5.3 Test Dataset 5.4 Quantitative Evaluation 6 Qualitative Analysis 7 Conclusion References Neural Networks Reduction via Lumping 1 Introduction 2 Related Work 3 Preliminaries 4 Lumping Neural Networks 5 Experimental Results 6 Conclusion References Knowledge Enhanced Neural Networks for Relational Domains 1 Introduction 2 Related Works 3 Knowledge Enhanced Neural Networks 3.1 KENN Architecture 3.2 Extending KENN for Relational Domains 3.3 Time Complexity 3.4 Treatment of Dependencies Among the Rules 4 Evaluation of the Model 4.1 Experimental Setup 4.2 Results 4.3 Comparison with Other NeSy Frameworks 5 Conclusions A Relational KENN Architecture B Results Distribution - Inductive Learning C Results Distribution - Transductive Learning D Comparison with SBR and RNM References Logic Tensor Networks for Top-N Recommendation 1 Introduction 2 Related Works 3 Background 3.1 Notation 3.2 Matrix Factorization 3.3 Logic Tensor Networks 3.4 Intuition of Real Logic Grounding 4 Method 4.1 Knowledge Base 4.2 Grounding of the Knowledge Base 4.3 Learning of the LTN 5 Experiments 5.1 Dataset 5.2 Experimental Setting 5.3 Evaluation Metrics 5.4 Training Details 6 Results 6.1 Training Time 7 Conclusions References Multiagent Systems A Review of the Muddy Children Problem 1 Introduction 2 The Muddy Children Problem 3 Properties 4 Possible Worlds Semantics 5 Formal Definition of the Muddy Children Problem 6 Related Literature and Existing Solutions 6.1 Eliminating Possible Worlds 6.2 Logic Programming 6.3 Other Potential Solutions 7 Variations of the Muddy Children Puzzle 8 Challenges 9 Conclusion References Multi-agent Cooperative Argumentation in Arg2P 1 Introduction 2 Background Notion: Structured Argumentation 3 Distributed Argumentation in Arg2P 3.1 Computational Model: The Master-Slave Actor Model 3.2 Implementation: The Parallel Library 4 Conclusion References Ethics by Design for Intelligent and Sustainable Adaptive Systems 1 Introduction 2 Ethics in Inductive Decision Systems 2.1 Ethics in Different Application Scenarios 2.2 Computational Methods for Fair Inductive Systems 2.3 Ethics, Principles and Truth-Makers 3 Formalizing Principles, Rules and Truth-Makers 4 Evaluating Ethical Compliance in Inductive Learning 4.1 Use Cases 4.2 Discussion of the Results 5 Conclusions References Automated Planning and Scheduling Verification of Numeric Planning Problems Through Domain Dynamic Consistency 1 Introduction 2 Background 2.1 Numeric Planning 2.2 Problem Relaxation and Heuristics 3 Domain Dynamic Consistency 4 Approximating Domain Dynamic Consistency 5 The Case of Urban Traffic Control 6 Conclusion References Comparing Multi-Agent Path Finding Algorithms in a Real Industrial Scenario 1 Introduction 2 Related Works 2.1 Single-Agent Path Finding 3 Problem Statement 3.1 Solutions 4 Experimental Evaluation 4.1 Implementation 4.2 Industrial Scenarios 4.3 Results 5 Conclusions References Logic-Based Ethical Planning 1 Introduction 2 Related Work 3 Model 3.1 LTL Language 3.2 Histories 3.3 Action Theory 3.4 Plans 3.5 Moral Conflicts 3.6 Lexicographic Value Base 3.7 Adding Desires 4 An Example 5 Computational Complexity 6 Conclusion References A Hybrid Recommender System with Implicit Feedbacks in Fashion Retail 1 Introduction 2 Collaborative Filtering with Implicit Feedbacks 3 System Architecture 3.1 Collaborative Filtering Module 3.2 Knowledge-Based Post-filtering Module 4 Case Study and Experimental Results 5 Conclusions and Related Works References Incremental Timeline-Based Planning for Efficient Plan Execution and Adaptation 1 Introduction 2 Related Works 3 Technical Background 3.1 A Lifted Heuristic for Timeline-Based Planning 3.2 An Explanatory Example 4 An Architecture for Deliberative and Reactive Reasoning 5 Plan Execution and Possible Adaptations 6 Experimental Results 7 Conclusions References Knowledge Acquisition and Completion for Long-Term Human-Robot Interactions Using Knowledge Graph Embedding 1 Introduction 2 Related Work 3 Methodology 3.1 Acquiring and Extending the Knowledge Base 3.2 Knowledge Graph Embedding 3.3 Long-Run 4 Results 5 Conclusions and Future Directions References Construct, Merge, Solve and Adapt Applied to a Bus Driver Scheduling Problem with Complex Break Constraints 1 Introduction 2 Problem Description 2.1 Problem Input 2.2 Solution 2.3 Work and Break Regulations 2.4 Objectives 3 Related Work 4 The CMSA Approach to the BDS Problem 4.1 The CMSA Algorithm 4.2 Greedy Heuristic 5 Experimental Results 5.1 Parameters Tuning 5.2 Analysis of the Results 6 Conclusions References Topic Modelling and Frame Identification for Political Arguments 1 Introduction 2 Related Work 3 Datasets 4 Topic Modeling and Frame Classification for Arguments 4.1 Generic Frame Classification 4.2 Topic Modeling and Issue-Specific Frame Identification 5 Results 6 Concluding Remarks References Substitute Plastic Film with Kraft Paper in Automatic Pallet Wrapping: An AI Pipeline 1 Motivations 1.1 A Note on Methodology and Purposes of This Work 2 Background 2.1 The ESTL Machine 3 Multiple Tracking of Bundles 4 Developing the Simulation 5 Experiments 6 Conclusions and Future Works References AI Applications Transformer Based Motion In-Betweening 1 Introduction 2 Related Work 3 Methodology 3.1 Tween Transformers (TWTR) 3.2 Loss Computation 3.3 Training 4 Setup and Experimental Results 4.1 Dataset 4.2 Evaluation Metrics 4.3 Data Preprocessing 4.4 Hyperparameters 4.5 Visualizer and Motion Generator 4.6 Inferences 5 Conclusion References A Logic-Based Tool for Dynamic Generation and Classification of Musical Content 1 Introduction 2 Combining Concepts: The Description Logic TCL 3 The Tool NERVOUS: Automated Generation of Prototypical Descriptions of Musical Genres 4 The Tool NERVOUS: Generation of Novel Musical Genres 5 The Tool NERVOUS: Reclassifications and Recommendations 6 Evaluation and Discussion 7 Conclusions and Future Works References Why Can Neural Networks Recognize Us by Our Finger Movements? 1 Introduction 2 Experimental Settings 2.1 Dataset 2.2 Application Architecture 2.3 Preprocessing Techniques 2.4 Neural Network Architecture 3 Results 4 Discussion References Miscellany Labelled Sequent Calculi for Conditional Logics: Conditional Excluded Middle and Conditional Modus Ponens Finally Together 1 Introduction 2 Conditional Logics with Selection Function Semantics 3 SeqS22: A Sequent Calculus for Conditional Logics 4 A Theorem Prover for Conditional Logics with CEM 5 Conclusions and Future Works References Deep Learning for ECoG Brain-Computer Interface: End-to-End vs. Hand-Crafted Features*-4pt 1 Introduction 2 Methods 2.1 Dataset 2.2 Data Representation and Problem 2.3 Hand-Crafted Features Extraction and DL Optimization 2.4 DL Architectures 2.5 Offline Experiments 3 Results 3.1 Filters Visualization 3.2 Target Perturbation 4 Discussion References Quantum Circuit Compilation for the Graph Coloring Problem 1 Introduction 2 Background 3 The QCC Problem 3.1 The Graph Coloring Problem 3.2 Quantum Gate Compilation Problem 4 A Greedy Randomized Search Algorithm 4.1 Compile Circuit Algorithm 5 Experimental Evaluation 5.1 Setup 5.2 Results 6 Conclusions References Toward a Heterogeneous Multi-robot Framework for Priority-Based Sanitization of Railway Stations 1 Introduction 2 The Architecture 2.1 Heatmap Definition and Update 2.2 Multi-agent Experience Replay and the Learning Process 3 Experiments 3.1 Case 1: The Training Process 3.2 Case 2: Testing with Real Data 4 Conclusions and Future Works References Simulated Annealing for the Home Healthcare Routing and Scheduling Problem 1 Introduction 2 Related Work 3 Problem Statement 4 Solution Technique 4.1 Search Space and Initial Solution 4.2 Neighborhood Operator 4.3 Simulated Annealing 5 Experimental Results 5.1 Instances 5.2 Parameter Tuning 5.3 Comparative Results 6 Conclusions References MAP Inference in Probabilistic Answer Set Programs 1 Introduction 2 Related Work 3 Background 4 MAP Inference in Probabilistic Answer Set Programming 4.1 Algorithm 5 Experiments 6 Conclusions References Verifying a Stochastic Model for the Spread of a SARS-CoV-2-Like Infection: Opportunities and Limitations 1 Introduction 2 Background 3 A Stochastic Model for SARS-CoV-2-Like Infection's Spread 4 Experimenting with State-of-the-art Stochastic Model Checkers 5 Related Works 6 Conclusions and Future Works References Natural Language Processing DelBERTo: A Deep Lightweight Transformer for Sentiment Analysis 1 Introduction 2 Background and Related Work 3 Proposed Method 3.1 Text Preprocessing 3.2 Training Procedure 4 Experiments 4.1 Datasets 4.2 Setup 4.3 Results 5 Conclusions and Discussion References A BERT-Based Scoring System for Workplace Safety Courses in Italian 1 Introduction 2 Background and Related Work 2.1 BERT 2.2 Related Work 3 A BERT-Based Scoring System 3.1 Dataset and Task Description 3.2 Approaches Proposed 3.3 BERT-Based Implementation 4 Experimental Results 4.1 Training Configurations 4.2 Approaches Comparison 5 Conclusion and Future Work References Embedding Contextual Information in Seq2seq Models for Grounded Semantic Role Labeling 1 Introduction 2 Related Work 3 Embedding Contextual Information in Seq2seq Models 3.1 Describing a Map in Natural Language 3.2 Embedding Frame Semantics 4 Experimental Evaluation 4.1 Testing Different Types of Input 5 Results 6 Conclusions References Keynote talk Adventures with Datalog: Walking the Thin Line Between Theory and Practice References Author Index