Progress in Artificial Intelligence 25th EPIA Conference on Artificial Intelligence, EPIA 2026, Madeira, Portugal, September 2–4, 2026, Proceedings, Part II

.- Natural Language Processing, Text Mining and Applications (NLP-TeMA).

.- From Prototype to Platform: SaiLR 2.0 for AI-Assisted Systematic Literature Reviews.

.- Separating Human from Machine: Deep Linguistic Analysis and Automatic Classification of Fake News in Brazilian Portuguese.

.- PorTEXTO: A European Portuguese Benchmark for Visual Text Extraction.

.- On the Robustness of Medical Image Report Generation: Diagnosis Label Entailment Under Perturbed Conditions.

.- NER Models for Portuguese Emergency Room Notes: Extracting Diagnoses, Medication Allergies, and Usual Medications.

.- Artificial Intelligence in Medicine (AIM).

.- Evaluating the Privacy–Utility Trade-off of Differential Privacy in Machine Learning Models for Clinical Outcome Prediction.

.- Assessing Cognitive Performance in Language Models: A Framework for Cognitive Impairment Simulation.

.- Musiquence AI Toolset: A Serious Game Ecosystem for Health Professionals.

.- Dealing with Long-Form Clinical Audio for Portuguese Veterinarians.

.- A Transition-Aware Framework for Wearable Sleepiness Scoring in Real-Time and Near-Real-Time Sleep Monitoring.

.- NeurAFTs: Neural Accelerated Failure Time Models.

.- Stability-Aware Graph Neural Networks For Functional Neuroimaging.

.- Beyond Downstream Accuracy: A Multi-Axis Framework for Evaluating Synthetic EEG in Seizure Detection.

.- Ethics and Responsibility in AI (ERAI).

.- Quality Assessment in Subgroup Discovery Using Uncertainty Targets: A Healthcare Use Case.

.- Grooming-like Patterns in AI Chatbot Interactions with Children: A Novel Perspective?.

.- Artificial Intelligence and Law (AIL).

.- Adversarial Attacks on Legal NLP: Evaluating the Robustness of Contract Analysis Models.

.- Human Oversight in AI Governance: Ensuring Effectiveness Through By-Design Measures and Coordinated Responsibilities.

.- Beyond Single-Step Judgments: Recursive Evaluation of Legal Summaries with Language Models.

.- VeLex: A Neuro-Symbolic Pipeline for Compliance-Matrix Extraction from Swiss Financial Regulation.

.- Gen AI for deceptive purposes in pre-trial investigations. Deepfakes vis-`a-vis human and fundamental rights.

.- Automatic Legal Descriptor Generation for Portuguese Legal Documents.

.- AI for Architecture, Engineering and Conservation (AI4AEC).

.- A Physics-Informed Neural Network for Muographic Density Estimation in Concrete.

.- Overlap-controlled frame extraction for underwater photogrammetry.

.- Adapting SAM2-UNet for High-Resolution LULC Semantic Segmentation with OpenStreetMap Auxiliary Supervision.

.- reBEN7-S2UNet: a Sentinel-2 U-Net for Seven-Class LULC Mapping in Portuguese Landscapes.

.- Spatial Agency Beyond the Prompt: An in-the-wild evaluation of Generative AI Literacy in Architectural Education.

.- Generative and Agentic Artificial Intelligence (GAAI).

.- Sensitivity Ridges in Conditioned Generative Models.

.- Large Language Models Are Not (Yet) Robust in Understanding Code Against Semantics-Preserving Mutations.

.- Agentic AutoML: LLM-driven Semantic Auditing for Model Integrity in Automated Feature Selection.

.- AMALIA-VL: A Native European Portuguese Open-Source Vision and Language Model.

.- Integrating Keypoint Modality into Large Language Model for Continuous Sign Language Translation.

.- On the Limits of Similarity-Based Policy Detection in Course-Specific RAG: A Bilingual Ablation Study.

.- Artificial Intelligence in Transportation Systems (AITS).

.- Robotic Navigation of a Delivery Robot Based on Behavior Trees.

.- A Robust Pipeline for Container Code Recognition Using YOLOv8 and TrOCR.

.- Generative Counterfactual Explanations for Electric Vehicle Eco-Driving.