Preface
Organization
Keynotes
Machine Learning Algorithms for Brain-Machine Interfaces
Digital Twins of the Ocean
On the Use (and Misuse) of Differential Privacy in Machine Learning
Learning on Graphs
Contents – Part I
Contents – Part II
Ambient Intelligence and Affective Environments
Simulation-Based Adaptive Interface for Personalized Learning of AI Fundamentals in Secondary School
1 Introduction
2 Education About AI
3 The RoboboITS
3.1 RoboboITS Architecture
3.2 RoboboITS Operation
4 AI Lesson Implemented
5 Secondary School Validation
6 Conclusions
References
Gamified CollectiveEyes: A Gamified Distributed Infrastructure for Collectively Sharing People’s Eyes
1 Introduction
2 Gamified CollectiveEyes
2.1 Seeing Several Viewpoints Simultaneously
2.2 Navigating Views with Gaze-Focused Gesture
2.3 Topic Channels
2.4 Thing-Focused and Value-Focused Topic Channel
2.5 Gamification Strategies in Gamified CollectiveEyes
3 A User Study for Motivation Management
3.1 Research Method
3.2 Effects of Topic Channels
3.3 Effects of Gamification
3.4 Effects of Consciousness
4 A User Study for Serendipity Management
4.1 Research Method
4.2 Effects of Serendipity
5 Related Work
6 Limitation of the Current Study
7 Conclusion and Future Work
References
Design and Development of Ontology for AI-Based Software Systems to Manage the Food Intake and Energy Consumption of Obesity, Diabetes and Tube Feeding Patients
1 Introduction
2 Related Works
2.1 FoodOn Ontology
2.2 Quisper Ontology
2.3 Ontology Based Food Recommendation
3 Methodology
3.1 Diabetes Use Case
3.2 Obesity Use Case
3.3 Tube Feeding Use Case
4 Proposed Ontology
5 Discussion
6 Conclusions
References
A System for Animal Health Monitoring and Emotions Detection
1 Introduction
2 Related Works
3 Methodology
4 System Overview
5 Experiments
6 Results Evaluation
7 Future Works
8 Conclusion
References
Ethics and Responsibility in Artificial Intelligence
A Three-Way Knot: Privacy, Fairness, and Predictive Performance Dynamics
1 Introduction
2 Background
2.1 Privacy
2.2 Fairness
2.3 Related Work
3 Experimental Study
3.1 Data
3.2 Methods
3.3 Experimental Results
4 Discussion
5 Conclusion
References
A Maturity Model for Industries and Organizations of All Types to Adopt Responsible AI—Preliminary Results
1 Introduction
1.1 Context and Justification
1.2 Why a Maturity Model for Responsible Artificial Intelligence (RAI)?
2 Methodology
3 The Maturity Model for Responsible AI
4 Implementation, Results and Discussion
5 Conclusions
References
Completeness of Datasets Documentation on ML/AI Repositories: An Empirical Investigation
1 Introduction and Motivation
2 Documentation Test Sheet from Related Works
2.1 Fields of Information
2.2 Measurement
3 Study Design
3.1 Repositories Under Analysis
3.2 Datasets Selection
4 Results and Discussion
4.1 Datasets Level
4.2 Sections Level
4.3 Test Fields Level
5 Threats to Validity and Limitations
6 Conclusions
7 Future Work
References
Navigating the Landscape of AI Ethics and Responsibility
1 Introduction
2 Research Methodology
3 Analysis of the Literature
4 Discussion
5 Conclusion
References
Towards Interpretability in Fintech Applications via Knowledge Augmentation
1 Introduction
2 Interpretability in Fintech
2.1 Interpretability Approaches
2.2 Surrogate Models
3 Knowledge Extraction and Augmentation
3.1 Knowledge Extraction Methods
3.2 Knowledge Augmentation Methods
4 Proposed Approach
5 Experimental Setup
5.1 Evaluation Metrics
5.2 Case Studies
6 Analysis of Results
7 Conclusions and Future Work
References
General Artificial Intelligence
Revisiting Deep Attention Recurrent Networks
1 Introduction
2 Related Work
2.1 Deep Attention Recurrent Q-Network
2.2 Soft Top-Down Spatial Attention
2.3 Similarities Between DARQN and STDA
3 Experimental Setup
3.1 Extensions to the DARQN Architecture
3.2 Top-Down Spatial Attention Agent
3.3 Training Setup
4 Experimental Results
4.1 Preliminary Results
4.2 Comparative Results (DARAC vs. TDA)
4.3 Visualization of the Attention Maps
4.4 Discussion
5 Conclusion
References
Pre-training with Augmentations for Efficient Transfer in Model-Based Reinforcement Learning
1 Introduction
2 Related Work
3 Method
3.1 Augmentation Scheme
3.2 Pre-training with Augmentations
4 Evaluation
4.1 Experimental Setup
4.2 Pre-training of Model-Based RL Agents
4.3 Atari Games
5 Conclusions
References
DyPrune: Dynamic Pruning Rates for Neural Networks
1 Introduction
2 Methodology
2.1 Dataset
2.2 Pruning Weights
2.3 Removing Neurons
3 Results
4 Discussion
5 Conclusions
References
Robustness Analysis of Machine Learning Models Using Domain-Specific Test Data Perturbation
1 Introduction
2 Literature Review
2.1 Image
2.2 Audio
2.3 Text
3 Experimental Setup
4 Results
5 Conclusion
References
Vocalization Features to Recognize Small Dolphin Species for Limited Datasets
1 Introduction and Related Work
2 Features
2.1 The Spectral Analysis Features
2.2 The Contour Analysis Features
3 Classification
3.1 Data
3.2 The Training Phase
4 Results and Discussion
5 Conclusion
References
Covariance Kernel Learning Schemes for Gaussian Process Based Prediction Using Markov Chain Monte Carlo
1 Introduction
2 Model
3 Empirical Illustration
3.1 Model for Univariate Case
3.2 Model for Multivariate Case
3.3 New Nonparametric Kernel
4 Results
5 Conclusion
References
Intelligent Robotics
A Review on Quadruped Manipulators
1 Introduction
2 Methodology
3 Quadruped Manipulators
3.1 Leg-Arm Approaches
3.2 Robotic Arm Addition
4 Motion Planning
4.1 Separate Systems (SS)
4.2 Combined Systems (CS)
4.3 Discussion
5 Kinematic Configuration
6 Conclusions
References
Knowledge Discovery and Business Intelligence
Pollution Emission Patterns of Transportation in Porto, Portugal Through Network Analysis
1 Introduction
2 Related Work
3 Data and Methods
3.1 Data Pre-processing
3.2 Road Transportation and Emission Network
4 Experimental Results
4.1 Emissions over Porto
4.2 Road Network Analysis
4.3 Discussion
5 Conclusion and Future Work
References
Analysis of Dam Natural Frequencies Using a Convolutional Neural Network
1 Introduction
2 Case Study: Cabril Dam
2.1 Dam Description
2.2 Continuous Vibration Monitoring System
3 Supervised Convolutional Neural Network (CNN) Proposed for the Analysis of Dam Natural Frequencies
3.1 Dataset
3.2 Main Model
3.3 CNN Hyperparameter Tuning
4 Results: Analysis of Natural Frequencies of Cabril Dam
5 Conclusion and Future Work
References
Imbalanced Regression Evaluation Under Uncertain Domain Preferences
1 Introduction
2 Imbalanced Regression
2.1 Relevance Functions
2.2 Evaluation
3 Sensitivity Evaluation and Relevance Uncertainty
4 Experimental Study
4.1 Methods
4.2 Results
5 Conclusions and Future Work
References
Studying the Impact of Sampling in Highly Frequent Time Series
1 Introduction
2 Related Work
3 Methodology
4 Experimental Setup
4.1 Algorithm
4.2 Datasets
4.3 Missing Data
4.4 Evaluation
5 Experiments
6 Results and Discussion
7 Conclusion and Future Work
References
Mining Causal Links Between TV Sports Content and Real-World Data
1 Introduction
2 Literature Review
3 Data
4 Methods
4.1 Granger Causality Test
4.2 Causal Analysis of TV Viewership in Liga NOS
5 Results and Discussion
6 Conclusion
References
Hybrid SkipAwareRec: A Streaming Music Recommendation System
1 Introduction
2 Related Work
2.1 Recommendations with Negative Implicit Feedback
2.2 Sequential Music Recommendation
3 Methodology and Proposed Solution
3.1 Action Set Generation
3.2 Next Best Action Recommendation
3.3 Next Best Items Recommendation
4 Experiments and Results
4.1 Data Setup and Model Training
4.2 Evaluation
5 Conclusions and Future Work
References
Interpreting What is Important: An Explainability Approach and Study on Feature Selection
1 Introduction
2 Related Works
3 Datasets and Methods
3.1 Rossmann Store Sales
3.2 Bike Sharing Dataset
3.3 Data Exploration
3.4 LSTM Hyperparameter Tunning
3.5 SHAP Method Implementation
4 Experiments, Results, and Discussion
4.1 Experimental Setup
4.2 Experimental Procedure
4.3 Results
4.4 Discussion
5 Conclusion and Future Work
References
Time-Series Pattern Verification in CNC Machining Data
1 Introduction
2 Background
2.1 CNC Machining and Offset Adjustment in Turning
2.2 Feature Extraction and Linear Frequency Cepstral Coefficients
2.3 One-Class Classification
3 Methodology
4 Results and Discussion
5 Conclusion
References
A Comparison of Automated Machine Learning Tools for Predicting Energy Building Consumption in Smart Cities
1 Introduction
2 Related Work
3 Materials and Methods
3.1 Data
3.2 AutoML Methods
3.3 Evaluation
4 Results
5 Conclusions
References
Measuring Latency-Accuracy Trade-Offs in Convolutional Neural Networks
1 Introduction
2 Related Work
3 Proposed Methodology
4 Results and Analysis
4.1 Case Study: Traffic Correlation
4.2 Results
4.3 Result Analysis
5 Conclusions and Future Work
References
MultiAgent Systems: Theory and Applications
Machine Learning Data Markets: Evaluating the Impact of Data Exchange on the Agent Learning Performance
1 Introduction
2 State of the Art
2.1 Multi-agent Systems and Machine Learning
2.2 Data Markets
3 Machine Learning Data Market
3.1 MLDM Architecture
3.2 Simple Negotiation Strategy
4 Experimental Result
4.1 The Experimental Setup
4.2 Results
5 Conclusion
References
Multi-robot Adaptive Sampling for Supervised Spatiotemporal Forecasting
1 Introduction
2 Problem Formulation
3 Proposed Approach
3.1 Spatiotemporal Mixture of Gaussian Processes (STMGP)
3.2 Mutual Information for Sampling
3.3 Training Local Gaussian Processes
4 Empirical Evaluation
4.1 Experimental Setup
4.2 Results
5 Concluding Remarks
References
Natural Language Processing, Text Mining and Applications
Topic Model with Contextual Outlier Handling: a Study on Electronic Invoice Product Descriptions
1 Introduction
2 Related Work
3 Topic Model with Contextual Outlier Handling
4 Case Study on Electronic Invoice Product Descriptions
5 Experimental Study
5.1 Experimental Setup
5.2 Obtained Results
6 Discussion
7 Conclusions
References
Tweet2Story: Extracting Narratives from Twitter
1 Introduction
2 Related Work
3 Tweet2Story: Automatic Extraction of Narratives
3.1 Identifying Narrative Elements and Relationships
3.2 Narrative Extraction Rules
3.3 Annotation Schema
4 Analysis of Results
4.1 As an OpenIE Tool
4.2 With Manually Annotated Data
4.3 Empirical Comparison of Visualizations
5 Conclusion
References
Argumentation Mining from Textual Documents Combining Deep Learning and Reasoning
1 Introduction
2 Related Work
3 The N-SAUR System
4 Experimental Setup
4.1 Data
4.2 Evaluation Methodologies
4.3 Experimental Results
5 Conclusions and Future Work
References
Event Extraction for Portuguese: A QA-Driven Approach Using ACE-2005
1 Introduction
2 Related Work
3 Methodology
3.1 Trigger Extraction
3.2 Argument Extraction
4 Data
4.1 ACE-2005 Translation
4.2 SQuAD Translation
5 Modeling
6 Results
7 Discussion
8 Conclusion
References
Symbolic Versus Deep Learning Techniques for Explainable Sentiment Analysis
1 Introduction
2 Related Work
3 Combining Symbolic and Deep Learning for SA
3.1 Corpus Used and its Annotation
3.2 Generation of Sentiment Lexicons
3.3 Shifter Patterns
3.4 Experimental Setup and Results
4 Conclusions
References
Assessing Good, Bad and Ugly Arguments Generated by ChatGPT: a New Dataset, its Methodology and Associated Tasks
1 Introduction
2 Background
2.1 Argument(ation) Mining
2.2 Automatic Essay Scoring
3 Generating Argumentative Essays with ChatGPT
4 ArGPT: Dataset Annotation and Statistics
5 Using ArGPT: Supported Tasks and Their Baselines
5.1 Evaluation Metrics
5.2 Results and Discussion
6 The Connection with Human Argumentation
7 Conclusions and Future Work
References
Advancing Neural Encoding of Portuguese with Transformer Albertina PT-
1 Introduction
2 Related Work
2.1 Encoders Whose Multilingual Data Set Included Portuguese
2.2 Encoders Specifically Concerned with Portuguese
3 Data Sets
3.1 Data Sets for the Pre-training Stage
3.2 Data Sets for the Fine-tuning Concerning Downstream Tasks
4 Albertina PT- Model
4.1 The Starting Encoder
4.2 Pre-training Albertina PT-BR
4.3 Pre-training Albertina PT-PT
4.4 Fine-tuning Albertina and BERTimbau
5 Experimental Results
5.1 Improving the State of the Art on ASSIN 2 Tasks
5.2 Setting the State of the Art on Portuguese GLUE Tasks
5.3 Discussion
6 Concluding Remarks
References
OSPT: European Portuguese Paraphrastic Dataset with Machine Translation
1 Introduction
2 Related Work
3 The Dataset
3.1 Choosing a Data Source
3.2 Automatic Quality Assessment, Cleaning, and Filtering
3.3 Data Analysis
4 Learning Sentence Embeddings
4.1 Experimental Setup
4.2 Results
5 Paraphrase Generation
5.1 Experimental Setup
5.2 Results
6 Conclusion
References
Task Conditioned BERT for Joint Intent Detection and Slot-Filling
1 Introduction
2 Related Work
3 Proposed Model
3.1 Dialogue Task Conditioned Encoder
3.2 Dialogue Task Conditioning
3.3 BERT-DST: Span Slots
3.4 BDST-I: Intent Detection
3.5 BDST-C: Categorical Slots
3.6 BDST-J: Joint Intent and Multiple-Slots
4 Evaluation
4.1 Datasets
4.2 Training
4.3 Metrics and Evaluation Methodology
4.4 General Results
5 Conclusion
References
Planning, Scheduling and Decision-Making in AI
Data-driven Single Machine Scheduling Minimizing Weighted Number of Tardy Jobs
1 Introduction
2 Solution Approach
3 Experimental Results
4 Conclusion
References
Heuristic Search Optimisation Using Planning and Curriculum Learning Techniques
1 Introduction
2 Related Work
3 Classical Planning
4 Planner's Architecture
5 The Proposed Neural Network
6 Curriculum Learning
7 Experimental Results
7.1 Training
7.2 Comparison to Prior State-of-the-Art
8 Conclusion and Future Work
References
Social Simulation and Modelling
Review of Agent-Based Evacuation Models in Python
1 Introduction
1.1 Evacuation
1.2 Agent-Based Modelling in Python
2 Systematic Review
2.1 Research Questions
2.2 Results
3 Conclusion
References
Author Index