Abstract
Automated seizure quantification and classification are needed for semiology-based epileptic seizure diagnosis support. To the best of our knowledge, the 5-class (Hypermotor, Automotor, Complex Motor, Psychogenic Non-Epileptic Seizures, and Generalized Tonic-Clonic Seizures) seizure video dataset (198 seizures from 74 patients) studied in this paper is the largest 5-class dataset ever curated, composed of monocular RGB videos from two university hospital epilepsy monitoring units. 2D skeletons were estimated using ViTPose, a vision transformer deep learning (DL) architecture, and lifted to 3D space using MotionBERT, a multimodal motion transformer architecture. The movements were quantified based on the estimated 3D skeleton sequences. Two approaches were evaluated for seizure classification: (1) classical machine learning methods (Random Forest (RF) and XGBoost) applied to quantified movement parameters, and (2) 2D skeleton-based DL using MotionBERT action, an action recognition DL model, to which we perform transfer-learning. The best model achieved a promising, above literature, 5-fold cross-validated macro average F1-score of 0.84 +/- 0.09 (RF) for 5-class classification. The binary case (Automotor vs Hypermotor) resulted in 0.80 +/- 0.18 (MotionBERT action), and adding a 3rd class (Complex motor) lowered to 0.65 +/- 0.14 (RF). This novel multi-stage classification ensures that the included movement features are traceable, allowing interpretable AI exploration of this novel approach supporting future clinical diagnosis.

Abstract
Federated Learning (FL) enables decentralized model training while preserving data privacy but remains susceptible to poisoning attacks. Malicious clients can manipulate local data or model updates, threatening FL's reliability, especially in privacy-sensitive domains like healthcare and finance. While client-side optimization algorithms play a crucial role in training local models, their resilience to such attacks is underexplored. This study empirically evaluates the robustness of three widely used optimization algorithms: SGD, Adam, and RMSProp-against label-flipping attacks (LFAs) in image classification tasks using Convolutional Neural Networks (CNNs). Through 900 individual runs in both federated and centralized learning (CL) settings, we analyze their performance under Independent and Identically Distributed (IID) and Non-IID data distributions. Results reveal that SGD is the most resilient, achieving the highest accuracy in 87% of cases, while Adam performs best in 13%. Additionally, centralized models outperform FL on CIFAR-10, whereas FL excels on Fashion-MNIST, highlighting the impact of dataset characteristics on adversarial robustness.

Abstract
Medical image captioning is an essential tool to produce descriptive text reports of medical images. One of the central problems of medical image captioning is their poor domain description generation because large pre-trained language models are primarily trained in non-medical text domains with different semantics of medical text. To overcome this limitation, we explore improvements in contrastive learning for X-ray images complemented with soft prompt engineering for medical image captioning and conditional text decoding for caption generation. The main objective is to develop a softprompt model to improve the accuracy and clinical relevance of the automatically generated captions while guaranteeing their complete linguistic accuracy without corrupting the models' performance. Experiments on the MIMIC-CXR and ROCO datasets showed that the inclusion of tailored soft-prompts improved accuracy and efficiency, while ensuring a more cohesive medical context for captions, aiding medical diagnosis and encouraging more accurate reporting.

Abstract
Group fairness in machine learning is an important area of research focused on achieving equitable outcomes across different groups defined by sensitive attributes such as race or gender. Federated learning, a decentralized approach to training machine learning models across multiple clients, amplifies the need for fairness methodologies due to its inherent heterogeneous data distributions that can exacerbate biases. The intersection of federated learning and group fairness has attracted significant interest, with 48 research works specifically dedicated to addressing this issue. However, no comprehensive survey has specifically focused on group fairness in Federated Learning. In this work, we analyze the key challenges of this topic, propose practices for its identification and benchmarking, and create a novel taxonomy based on criteria such as data partitioning, location, and strategy. Furthermore, we analyze broader concerns, review how different approaches handle the complexities of various sensitive attributes, examine common datasets and applications, and discuss the ethical, legal, and policy implications of group fairness in FL. We conclude by highlighting key areas for future research, emphasizing the need for more methods to address the complexities of achieving group fairness in federated systems.

Abstract
PurposeDetecting anomalous access to electronic health records (EHRs) is critical for safeguarding patient privacy and ensuring compliance with healthcare regulations. Traditional anomaly detection methods often struggle in this domain due to extreme class imbalance, limited labelled data and the subtlety of insider threats. This study proposes a lightweight, hybrid anomaly detection framework that integrates unsupervised, supervised and rule-based approaches using a meta-classifier architecture.Design/methodology/approachAn experimental and model-development approach is employed, combining machine learning techniques with domain-inspired rule modelling to construct a hybrid anomaly detection framework for healthcare access logs. Performance of the algorithm is measured using standard classification metrics such as precision, recall, F1-score and accuracy.FindingsEvaluated on a synthetic but realistic dataset of 50.000 normal and 500 labelled anomalous healthcare access events, the proposed framework achieved superior performance compared to standalone models as well as other hybrid models, with an F1-score of 0.8989 and recall of 0.8180. It also maintained low inference latency (0.028 ms) and energy consumption (4.03e-07 kg CO2), making it suitable for deployment in resource-constrained clinical environments.Originality/valueThis study highlights the potential of a hybrid meta-classifier to enhance anomaly detection in healthcare access logs, capturing both subtle and obvious anomalies while outperforming conventional models and remaining efficient, scalable and practical for real-time monitoring.

Abstract
Humour poses a unique challenge for artificial intelligence, as it often relies on non-literal language, cultural references, and linguistic creativity. The JOKER Lab, now in its fourth year, aims to advance computational humour research through shared tasks on curated, multilingual datasets, with applications in education, computer-mediated communication and translation, and conversational AI. This paper provides an overview of the JOKER Lab held at CLEF 2025, detailing the setup and results of its three main tasks: (1) humour-aware information retrieval, which involves searching a document collection for humorous texts relevant to user queries in either English or Portuguese; (2) pun translation, focussed on humour-preserving translation of paronomastic jokes from English into French; and (3) onomastic wordplay translation, a task addressing the translation of name-based wordplay from English into French. The 2025 edition builds upon previous iterations by expanding datasets and emphasising nuanced, manual evaluation methods. The Task 1 results show a marked improvement this year, apparently due to participants' judicious combination of retrieval and filtering techniques. Tasks 2 and 3 remain challenging, not only in terms of system performance but also in terms of defining meaningful and reliable evaluation metrics.