Artificial Intelligence and Cognitive Science: 33rd Irish Conference, AICS 2025, Dublin, Ireland, December 1–2, 2025, Revised Selected Papers(2950 Communications in Computer and Information Science)

This book constitutes the proceedings of the 33rd Irish Conference on Artificial Intelligence and Cognitive Science, AICS 2025, held in Dublin, Ireland, during December 1–2, 2025.

The 33 full papers presented in these proceedings were carefully reviewed and selected from 122 submissions. These papers cover themes such as applications of AI, machine learning and deep learning methods, dataset construction and analysis, AI for social good, AI ethics and responsible AI, cognitive science and learning theory, and semantic web technologies. They reflect both theoretical advances and practical applications of AI technologies in different domains.

.- Architecture-Dependent Synthetic Data Strategies for Leukemia Diagnosis with Explainable Deep Learning.
.- AI-Generated Images: What Humans and Machines See When They Look at the Same Image.
.- Toward Trustworthy Policy Discovery: A Multilingual Hybrid RAG Approach Applied to EU and UK Peatland Restoration.
.- Towards generating more interpretable counterfactuals via concept vectors: a preliminary study on chest X-rays.
.- Fault Isolation & Explainable AI for Satellite Health Monitoring.
.- Bitcoin Prediction: A Hybrid Approach with Recurrent Neural Networks and GPT-based Reasoning.
.- Evaluating Pre-trained 3rd Party AI Models in Medical Device Software Development for Responsible and Ethical Use.
.- Enhancing Healthcare Data Transparency through a FAIR Ontology for the Systems Engineering Initiative for Patient Safety (SEIPS).
.- Speech Emotion Recognition from Audio using Machine Learning and Deep Learning Approaches.
.- Hindsight Preference Replay Improves Preference-Conditioned Multi-Objective Reinforcement Learning.
.- Classifying Decision-Based Effector Modalities from Fragmented iEEG via Dynamic Recursive Feature Elimination.
.- An Optimized Deep Learning-Based Structure-from-Motion Pipeline for Efficient Glacier 3D Reconstruction.
.- Formal Robustness Verification of a Learnt Autonomous Driver.
.- In Machina N400: Pinpointing Where a Causal Language Model Detects Semantic Violations.
.- TrailSurfNET: Automated Hiking Trail Surface Classification Using CNNs and OpenStreetMap Annotations.
.- LFE-SCCM: Low-Frequency Enhanced Method for Screen Content Image Compression.
.- CAViT - Channel-Aware Vision Transformer for Dynamic Feature Fusion.
.- LLM-RAG Retrieval Optimisation through Fixed and Adaptive Parameters with Cumulative Thresholds.
.- Cooperative MARL with Structured Rewards and Explainable Agents for Hyperparameter Tuning.
.- A Continuous Cloud Coverage Dataset for Robust AI in Occluded Environments.
.- Forecasting Energy Demand in Buildings: The Case for Trees over Deep Nets.
.- Evaluating the Impact of Situation Testing on SMOTE-Based Sampling Techniques for Bias Mitigation.
.- Modality Fusion of MRI and Clinical Data for Glioma Tumour Grading.
.- Tailored Loss Functions to Improve Vertebra Centroid Localization and Classification in Sagittal Spinal Radiographs.
.- Gaussian Process-based Motor Imagery Classification.
.- Synthetic Data for PET and HDPE Classification: A Comparison of Diffusion and Autoregressive Models.
.- Improving Low-Resource Language Translation via Targeted Vocabulary Expansion.
.- Profit-Sensitive Benchmarking of Post-hoc Explainability in Customer Churn.
.- Latency–Sustainability Trade-offs in On-Device vs. Near-Edge Offloaded Semantic Segmentation.
.- Operationalising Properties of Complex Adaptive Systems for ECG-Based Emotion Recognition.
.- Evaluating the Efficacy of Monolingual and Multilingual Pre-training for Igbo Automatic Speech Recognition.
.- A Novel Deep Learning Framework for DeepFake Detection Using Attention-Enhanced EfficientNet-B7.