publications

Large Language Models (LLMs) are increasingly proposed as near-autonomous artificial intelligence (AI) agents capable of making everyday decisions on behalf of humans. Although LLMs perform well on many technical tasks, their behavior in personal decision-making remains less understood. Previous studies have assessed their rationality and moral alignment with human decisions. However, the behavior of AI assistants in scenarios where financial rewards are at odds with user comfort has not yet been thoroughly explored. In this paper, we tackle this problem by quantifying the prices assigned by multiple LLMs to a series of user discomforts: additional walking, waiting, hunger, and pain. We uncover several key concerns that strongly question the prospect of using current LLMs as decision-making assistants: (1) a large variance in responses between LLMs, (2) within a single LLM, responses show fragility to minor variations in prompt phrasing (e.g., reformulating the question in the first person can considerably alter the decision), (3) LLMs can accept unreasonably low rewards for major inconveniences (e.g., 1 euro to wait 10 hours), (4) LLMs can reject monetary gains where no discomfort is imposed (e.g., 1,000 euro to wait 0 minutes), (5) models often make different decisions at rounded figures (e.g., 10 euro, 100 euro, or 1,000 euro), and (6) smaller models fail to demonstrate consistent decision-making behavior. These findings emphasize the need for scrutiny of how LLMs value human inconvenience, particularly as we move toward applications where such cash-versus-comfort trade-offs are made on users’ behalf.

Artificial intelligence models are increasingly scaled to improve predictive accuracy, yet it remains unclear whether scale improves the quality of post-hoc explanations. We investigate this relationship by evaluating 11 computer vision models representing increasing levels of depth and complexity within the ResNet, DenseNet, and Vision Transformer families, trained from scratch or pretrained, across the medical and natural image datasets with ground-truth segmentation masks. For each model, we generate explanations using five post-hoc explainable AI methods and quantify mask alignment using two localisation metrics: Relevance Rank Accuracy and the proposed Dual-Polarity Precision, which measures positive attributions inside the class mask and negative attributions outside it. Across datasets and methods, increasing architectural depth and parameter count does not improve explanation quality in most statistical comparisons, and smaller models often match or exceed deeper variants. While pretraining typically improves predictive performance and increases the dependence of explanations on learned weights, it does not consistently increase localisation scores. We also observe scenarios in which models achieve strong predictive performance while localisation precision is near zero, suggesting that performance metrics alone may not indicate whether predictions are based on the annotated regions. These results indicate that larger models do not reliably provide higher-quality explanations, and that explainability should therefore be assessed explicitly during model selection for safety-sensitive deployments.

Explainable AI (XAI) aims to make the behaviour of machine learning models interpretable, yet many explanation methods remain difficult to understand. The integration of Natural Language Generation into XAI aims to deliver explanations in textual form, making them more accessible to practitioners. Current approaches, however, largely yield static lists of feature importances. Although such explanations indicate what influences the prediction, they do not explain why the prediction occurs. In this study, we draw on insights from social sciences and linguistics, and argue that XAI explanations should be presented in the form of narratives. Narrative explanations support human understanding through four defining properties: continuous structure, cause-effect mechanisms, linguistic fluency, and lexical diversity. We show that standard Natural Language Processing (NLP) metrics based solely on token probability or word frequency fail to capture these properties and can be matched or exceeded by tautological text that conveys no explanatory content. To address this issue, we propose seven automatic metrics that quantify the narrative quality of explanations along the four identified dimensions. We benchmark current state-of-the-art explanation generation methods on six datasets and show that the proposed metrics separate descriptive from narrative explanations more reliably than standard NLP metrics. Finally, to further advance the field, we propose a set of problem-agnostic XAI Narrative generation rules for producing natural language XAI explanations, so that the resulting XAI Narratives exhibit stronger narrative properties and align with the findings from the linguistic and social science literature.

Background: Glaucoma is a progressive neurodegenerative disease of the optic nerve and one of the leading causes of irreversible blindness worldwide. Small RNAs (including miRNAs) play an important role in the pathogenesis of the disease. Despite extensive research, the exact mechanisms underlying glaucoma remain incompletely understood and current diagnostic tools primarily detect the disease at later stages, when substantial neuronal loss has already occurred. This study aimed to comprehensively analyze small RNA expression in the aqueous humor of patients with glaucomatous disease and identify specific RNA biomarkers associated with disease progression. Materials and Methods: This prospective, single-center, cross-sectional study involved 51 individuals: 31 adult patients with primary open-angle glaucoma (POAG) and 20 patients with cataracts qualified for elective surgery. A gradient-boosting decision tree classifier was implemented to classify POAG from cataracts. The model performance was rigorously assessed during cross-validation using a suite of metrics, including receiver operating characteristic (ROC) curves, area under the ROC curve, accuracy, recall, sensitivity, F1 score, specificity, precision, and negative predictive value. Results: Expression of hsa-miR-21-5p was significantly higher in POAG than that in cataracts, indicating its potential as both a biomarker and therapeutic target. Moreover, miR-210-3p expression was associated with apoptosis occurrence in trabecular meshwork cells treated with TGF-β1. Conclusions: Our study highlights the diagnostic potential of small-RNAs for distinguishing POAG from cataracts by integrating molecular profiling with advanced bioinformatics approaches. These findings provide new avenues for early diagnosis, personalized medicine, and improved clinical management.

Graph Neural Networks (GNNs) are a powerful technique for machine learning on graph-structured data, yet they pose challenges in interpretability. Existing GNN explanation methods usually yield technical outputs, such as subgraphs and feature importance scores, that are difficult to understand, thereby violating the purpose of explanations. Motivated by recent Explainable AI (XAI) research, we propose GraphXAIN, a method that generates natural language narratives explaining GNN prediction. GraphXAIN is a model- and explainer-agnostic method that uses Large Language Models (LLMs) to translate explanatory subgraphs and feature importance scores into coherent, story-like explanations of GNN predictions. Evaluation on real-world datasets demonstrates GraphXAIN’s ability to improve graph explanations. A survey of machine learning researchers and practitioners reveals that GraphXAIN enhances four explainability dimensions: understandability, satisfaction, convincingness, and suitability for communicating model predictions. When combined with another graph explainer method, GraphXAIN further improves trustworthiness, insightfulness, confidence, and usability. Notably, 95% of participants find GraphXAIN to be a valuable addition to the current GNN explanation methods. By incorporating natural language narratives, our approach serves both graph practitioners and non-expert users by providing clearer and more effective explanations.