AIxIA 2020 - Advances in Artificial Intelligence XIXth International Conference of the Italian Association for Artificial Intelligence, Virtual Event, November 25-27, 2020, Revised Selected Papers. Lecture Notes in Artificial Intelligence

This book constitutes the refereed post proceedings of the XIXth International Conference of the Italian Association for Artificial Intelligence, AIxIA 2020, held in Milano, Italy, in November 2020. Due to the COVID-19 pandemic, the conference was "rebooted"/ re-organized w.r.t. the original format. The 27 full papers were carefully reviewed and selected from 89 submissions. The society aims at increasing the public awareness of Artificial Intelligence, encouraging the teaching and promoting research in the field. Preface Organization Contents Explainable Artificial Intelligence Exploring Contextual Importance and Utility in Explaining Affect Detection 1 Introduction 2 Background 3 Contextual Importance and Contextual Utility 4 Dataset Description and Preprocessing 5 Outcome Explanations 5.1 Generated Explanation on WESAD 5.2 Generated Explanation on MAHNOB-HCI 6 Conclusion and Future Works References Explainable and Ethical AI: A Perspective on Argumentation and Logic Programming 1 Introduction 2 Explanation: Meaning and Roles 3 Logic Techniques for XAI 3.1 Why Logic? 3.2 User Requirements for XAI 3.3 Logic Approaches and Technologies Involved for XAI 4 System Architecture 4.1 Main Enabling Technologies 5 Preliminary Investigation: Examples 6 Related Works and Discussion 7 Conclusion References Understanding Automatic Pneumonia Classification Using Chest X-Ray Images 1 Introduction 2 Related Work 3 Proposed Approach 3.1 Classification 3.2 Visual Explanations 4 Experimental Protocol 4.1 Dataset Description 4.2 Training Phase 4.3 Performance Metrics 5 Results and Discussion 5.1 Classification Performance 5.2 Assessing Explanations from GradCAM 6 Conclusion References SeXAI: A Semantic Explainable Artificial Intelligence Framework 1 Introduction 2 Related Work 3 The Framework 4 SeXAI in Action 4.1 Quantitative Evaluation 4.2 Discussion 5 Conclusions References Explainable Attentional Neural Recommendations for Personalized Social Learning 1 Introduction 2 WhoTeach 3 Main Concepts and Definitions 4 GAT Models 5 Numerical Experiments 6 Expected Results for XAM 7 Conclusions and Future Works References Evolutionary and Population-Based Evolutionary Optimization of Graphs with GraphEA 1 Introduction and Related Works 2 GraphEA 2.1 Representation 2.2 Evolutionary Algorithm 3 Experimental Evaluation 3.1 Regression 3.2 Text Extraction 3.3 Robotic Controller Optimization 4 Concluding Remarks References Where the Local Search Affects Best in an Immune Algorithm 1 Introduction 2 The Hybrid Immune Algorithm 2.1 Hybrid-IA for the Community Detection 3 Experimental Results 3.1 Functional Sensitivity Analysis 4 Conclusions References Advances in Argumentation in Artificial Intelligence An Efficient Algorithm for Semi-stable Extensions 1 Introduction 2 Argumentation Framework and Complete Labelling Encodings 3 Overview of AASExts 4 Experimental Assessment of Encodings 5 Evaluation of AASExts 5.1 Experimental Setup 5.2 Experimental Results 6 Conclusion References Introducing General Argumentation Frameworks and Their Use 1 Introduction 2 Basics and Related Work 3 The Generalized Argumentation Framework 3.1 Mapping from Existing Frameworks 3.2 Mapping to Existing Frameworks 4 Sample Extensions of GAFs 4.1 Adding User and Topic Information 4.2 Expressing Evidence and Necessity 5 Matrix Representation and Examples 6 Conclusion References Towards an Implementation of a Concurrent Language for Argumentation 1 Introduction 2 Abstract Argumentation Frameworks 3 A Four-State Labelling Semantics 4 A Concurrent Argumentation Language 4.1 Semantics of Failure 4.2 Belief Revision and the AGM Framework 5 Implementation 6 Related Work 7 Conclusion and Future Work References Planning and Scheduling A Fault-Tolerant Automated Flight Path Planning System for an Ultralight Aircraft 1 Introduction 1.1 Related Work 1.2 System Description 2 Flight Path Planning System 2.1 Flight Path Planning in PDDL+ 2.2 Trajectory Planning Under Non-nominal Conditions 3 System Validation and Results 3.1 Test Scenarios 3.2 Test Results and Discussion 4 Conclusion and Future Work References In Defence of Design Patterns for AI Planning Knowledge Models 1 Introduction 2 A Knowledge Engineering (Historical) Perspective 3 Beyond Knowledge Encoding: Explainability 4 Example: The Mobile Design Pattern 5 Functionalities of the Centralised Repository 6 Discussion on Languages for Planning Knowledge Models 7 Conclusion References Solving Operating Room Scheduling Problems with Surgical Teams via Answer Set Programming 1 Introduction 2 Answer Set Programming 3 Problem Description 3.1 Informal Description 3.2 Mathematical Formulation 4 ASP Encoding for ORS with Surgical Teams 4.1 Data Model 4.2 Encoding 5 Experimental Results 5.1 Benchmarks 5.2 Results 5.3 Extended Analysis 6 Related Work 7 Conclusions References Artificial Intelligence and Robotics Optimal Control of Point-to-Point Navigation in Turbulent Time Dependent Flows Using Reinforcement Learning 1 Introduction 2 Methods 3 Results (Time-Independent Flows) 3.1 Shortest Time, No Energy Constraints 3.2 Minimal Energy Consumption 4 Results (Time-Dependent Flow) 5 Conclusions References Brain-Driven Telepresence Robots: A Fusion of User's Commands with Robot's Intelligence 1 Introduction 2 Taxonomy of Shared Approaches 3 Shared Autonomy Approaches Based on Pre-coded Behaviors 4 Shared Intelligence Approach Based on Policies 5 ROS-Neuro: A Common Framework for Developing Neurorobotics Applications 6 Conclusion References Knowledge-Driven Conversation for Social Robots: Exploring Crowdsourcing Mechanisms for Improving the System Capabilities 1 Introduction 2 Knowledge-Driven Conversation 3 Crowdsourcing Mechanisms for the Run-Time Expansion of the Knowledge Base 4 Recognition of Relevant Concepts 4.1 Dialogflow: Agents and Intents 4.2 Dialogflow Agent Training 5 Experiment and Discussion 6 Conclusions References Natural Language for Artificial Intelligence Grounding Dialogue History: Strengths and Weaknesses of Pre-trained Transformers 1 Introduction 2 Related Work 3 Dataset 4 Models 4.1 Language-Only Encoders 4.2 Multimodal Encoders 5 Experiments 5.1 Task Success 5.2 Are Models Sensitive to the Strategy Seen During Training? 5.3 The Role of the Last Question 5.4 How Attention Is Distributed Across Turns 5.5 Qualitative Evaluation 5.6 Details for Reproducibility 6 Conclusion References Breaking Down High-Level Robot Path-Finding Abstractions in Natural Language Programming 1 Introduction 2 Previous Work 3 Problem Formulation and System Design 3.1 High-Level to Low-Level (HL2LL) System Overview 3.2 Map Representation 3.3 Lee's Algorithm and Its Adaption 3.4 Path Information Extraction for NLPr 3.5 Path to Low-Level Sentence 3.6 Generating Code from NL Specifications Using the NLPr System 3.7 Formal Validation Using HFSM 4 Experimental Results 4.1 Case Study 5 Future Work 6 Conclusion References Experimental Evaluation of Algorithms for Solving Problems with Combinatorial Explosion Interleaving Levels of Consistency Enforcement for Singleton Arc Consistency in CSPs, with a New Best (N)SAC Algorithm 1 Introduction 2 Background 2.1 General Concepts 2.2 (N)SAC Algorithms 3 To Interleave or Not: Some (N)SAC Variants 4 (N)SAC with and Without AC: Initial Experiments 5 (N)SAC with and Without AC: More Extended Tests 6 Comparisons with an AC4-Style Interleaving Algorithm 7 Other Kinds of Interleaving 8 Conclusions References Improving the Efficiency of Euclidean TSP Solving in Constraint Programming by Predicting Effective Nocrossing Constraints 1 Introduction 2 Preliminaries 3 Avoiding Crossings 4 The Collected Data 5 Machine Learning the Goodness of Constraint Propagators 6 Experiments 6.1 Results of the Machine Learning Step 6.2 Results of the Overall Euclidean TSP Solver 7 Related Work 8 Conclusions and Future Work References From Contrastive to Abductive Explanations and Back Again 1 Introduction 2 Preliminaries 3 Contrastive vs. Abductive Explanations 3.1 Defining Abductive Explanations (AXps) 3.2 Defining Contrastive Explanations (CXps) 3.3 Relating Abductive and Contrastive Explanations 4 Extracting and Enumerating Explanations 5 Experimental Evaluation 5.1 Enumeration of CXps 5.2 Enumeration of CXps and AXps 6 Conclusions References Artificial Intelligence for an Ageing Society Towards Positive Artificial Intelligence 1 Introduction 2 Positive Psychology 3 Positive Technologies 4 Intelligent Agents 5 Positive Artificial Intelligence 5.1 Illustrative Scenario I: Improvements in Walking Experience for Elderly Pedestrians in Urban Roads 5.2 Illustrative Scenario II: Extended Interaction with Medical Devices for Treatment of Noncommunicable Diseases 6 Conclusion References Management at the Edge of Situation Awareness During Patient Telemonitoring 1 Introduction 2 Background and Related Work 3 Edge Computing Cognitive Architecture 3.1 System Architecture 3.2 CPAC: The Clinical Pathway Adherence Checker Module 3.3 CPAD: The Clinical Pathway Anomaly Detection Module 4 SARS-CoV-2 Patient Monitoring Scenario 5 Concluding Remarks References A Community-Based Activity Center to Promote Social Engagement and Counteract Decline of Elders Living Independently 1 Introduction 2 A Community-Based Activity Center 2.1 The CBAC Architecture 2.2 The Client Side 2.3 The Server Side 2.4 The User Side 2.5 Activities 2.6 Data Logged by the Platform 2.7 Recommendations 2.8 Additional CBAC Functionalities 3 Experimental Evaluation 3.1 Pre-pilot 3.2 Final Pilot 4 Discussion 5 Conclusion References Age-Related Walkability Assessment: A Preliminary Study Based on the EMG 1 Introduction 2 Experimental Setup for Data Collection 3 EMG Data Analysis 3.1 Subject Based Preprocessing 3.2 EMG Features Extraction 4 Experimental Results 4.1 Forced Speed Walk 4.2 Free Walk 4.3 Collision Avoidance 5 Conclusions References Discussion Papers It's the End of the Gold Standard as We Know It 1 Introduction 2 A Quick Primer on Linguistic Annotation 3 The Annotation of Highly Subjective Phenomena 4 The Power of Non-aggregated Data 5 Perspective-Aware Evaluation 6 An Experiment with Synthetic Data 7 Perspective-Aware Models and Impact on ExplainabilityA warning to the reader: this section contains strong language, in particular toward the end, merely employed for the sake of the exposition of experimental results. 8 Conclusion References Popularize Artificial Intelligence-“Pietro Torasso” Award Shared Intelligence for User-Supervised Robots: From User's Commands to Robot's Actions 1 Introduction 2 Methods 2.1 Brain-Computer Interface System to Detect the User's Commands 2.2 The Fusion of the User's Commands with the Robot's Intelligence 3 Experiments 4 Preliminary Results 5 Conclusion References Author Index