Vision-Language Models (VLMs) exhibit puzzling failures in multi-object visual tasks, such as hallucinating non-existent elements or failing to identify the most similar objects among distractions. While these errors mirror human cognitive constraints, such as the "Binding Problem", the internal mechanisms driving them in artificial systems remain poorly understood. Here, we propose a mechanistic insight by analyzing the representational geometry of open-weight VLMs (Qwen, InternVL, Gemma), comparing methodologies to distill "concept vectors" - latent directions encoding visual concepts. We validate our concept vectors via steering interventions that reliably manipulate model behavior in both simplified and naturalistic vision tasks (e.g., forcing the model to perceive a red flower as blue). We observe that the geometric overlap between these vectors strongly correlates with specific error patterns, offering a grounded quantitative framework to understand how internal representations shape model behavior and drive visual failures.

Graph eXplainable AI (G-XAI) is increasingly important for making Graph Neural Networks interpretable and accountable. While a growing number of explainers are available, choosing the right method and assessing the trustworthiness of its outputs remains unclear. Consistent evaluation practices and actionable guidance are still missing, hindering practical adoption. In this paper, we introduce a unified, quantitative benchmarking framework for G-XAI that requires no ground-truth assumptions. We formalize tabular explainability metrics for graph data, evaluating topological structure and node features as independent components. Our large-scale benchmarking study identifies explainers that consistently lie on the Pareto front across metric pairs and tasks, establishing robustly non-dominated solutions - while confirming that no single explainer achieves universal superiority. We distill our findings into actionable G-XAI usability guidelines to support Machine Learning practitioners in evaluating and deploying trustworthy GNN-based pipelines.

Agent-Based Models (ABMs) are powerful tools for studying emergent properties in complex systems. In ABMs, agent behaviors are governed by local interactions and stochastic rules. However, these rules are, in general, non-differentiable, limiting the use of gradient-based methods for optimization, and thus integration with real-world data. We propose a novel framework to learn a differentiable surrogate of any ABM by observing its generated data. Our method combines diffusion models to capture behavioral stochasticity and graph neural networks to model agent interactions. Distinct from prior surrogate approaches, our method introduces a fundamental shift: rather than approximating system-level outputs, it models individual agent behavior directly, preserving the decentralized, bottom-up dynamics that define ABMs. We validate our approach on two ABMs (Schelling's segregation model and a Predator-Prey ecosystem) showing that it replicates individual-level patterns and accurately forecasts emergent dynamics beyond training. Our results demonstrate the potential of combining diffusion models and graph learning for data-driven ABM simulation.

Determining whether an algorithmic decision-making system discriminates against a specific demographic typically involves comparing a single point estimate of a fairness metric against a predefined threshold. This practice is statistically brittle: it ignores sampling error and treats small demographic subgroups the same as large ones. The problem intensifies in intersectional analyses, where multiple sensitive attributes are considered jointly, giving rise to a larger number of smaller groups. As these groups become more granular, the data representing them becomes too sparse for reliable estimation, and fairness metrics yield excessively wide confidence intervals, precluding meaningful conclusions about potential unfair treatments. In this paper, we introduce a unified, size-adaptive, hypothesis‑testing framework that turns fairness assessment into an evidence‑based statistical decision. Our contribution is twofold. (i) For sufficiently large subgroups, we prove a Central‑Limit result for the statistical parity difference, leading to analytic confidence intervals and a Wald test whose type‑I (false positive) error is guaranteed at level $\alpha$. (ii) For the long tail of small intersectional groups, we derive a fully Bayesian Dirichlet–multinomial estimator; Monte-Carlo credible intervals are calibrated for any sample size and naturally converge to Wald intervals as more data becomes available. We validate our approach empirically on benchmark datasets, demonstrating how our tests provide interpretable, statistically rigorous decisions under varying degrees of data availability and intersectionality.

As deep learning systems become pervasive, the demand for trustworthy and transparent AI continues to grow. Traditional feature attribution methods, however, often lack robustness and alignment with human reasoning. This tutorial moves beyond feature attribution by introducing participants to two complementary interpretability paradigms: Concept-Based Explainable AI (C-XAI) and Mechanistic Interpretability. C-XAI provides explanations grounded in high-level, human-interpretable concepts, bridging the gap between model reasoning and human understanding. In parallel, mechanistic interpretability--a quickly emerging field--focuses on reverse-engineering neural networks to uncover and disentangle the internal mechanisms that give rise to human-understandable representations. Through interactive coding sessions and hands-on exercises, attendees will gain practical experience implementing, evaluating, and comparing a variety of C-XAI and mechanistic interpretability techniques. By the end of the tutorial, participants will be equipped with a modern interpretability toolbox and a deeper understanding of how to apply them in real-world scenarios.

Modern Machine Learning (ML) has significantly advanced various fields; yet, the challenge of understanding complex models, often referred to as the "black box problem", remains a barrier to their widespread adoption, particularly in critical domains such as medical diagnosis and financial services. Explainable AI (XAI) addresses this challenge by augmenting ML models' outputs with interpretable information to facilitate human understanding of their internal decision processes. Despite the proliferation of explainers in recent years, covering a wide range of ML tasks and explanation types, there is no consensus on what constitutes a good explanation, leaving ML practitioners without clear guidance for selecting appropriate explainers. We argue that explanation quality quantification is the enabling factor for informed explainer choices, but many proposed explanation evaluation criteria are either narrow in scope or closer to desired properties than quantifiable metrics. This paper addresses this gap by proposing a standardized set of metrics for quantitatively evaluating explanations across diverse explanation types and ML tasks. We describe in detail the metrics of Effective Compactness, Rank Quality Index and Stability, designed to assess quantitatively explanation quality for various types of explanations (attributions, counterfactuals and rules) across different ML tasks (classification, regression and anomaly detection). We then present an exhaustive benchmarking framework for tabular-based ML, comprising open datasets, trained models, and state-of-the-art explainers. For each (data, model, explainer) tuple, we measure the time of the explanation production, apply our metrics and collect the results, highlighting correlations and trade-offs between desired properties. The resulting framework allows us to quantitatively rank explainers suitable for specific ML scenarios and select the most appropriate one based on the user's requirements.

Modern Machine Learning (ML) has significantly advanced various research fields, but the opaque nature of ML models hinders their adoption in several domains. Explainable AI (XAI) addresses this challenge by providing additional information to help users understand the internal decision-making process of ML models. In the field of neuroscience, enriching a ML model for brain decoding with attribution-based XAI techniques means being able to highlight which brain areas correlate with the task at hand, thus offering valuable insights to domain experts. In this paper, we analyze human and Computer Vision (CV) systems in parallel, training and explaining two ML models based respectively on functional Magnetic Resonance Imaging (fMRI) and movie frames. We do so by leveraging the "StudyForrest" dataset, which includes functional Magnetic Resonance Imaging (fMRI) scans of subjects watching the "Forrest Gump" movie, emotion annotations, and eye-tracking data. For human vision the ML task is to link fMRI data with emotional annotations, and the explanations highlight the brain regions strongly correlated with the label. On the other hand, for computer vision, the input data is movie frames, and the explanations are pixel-level heatmaps. We cross-analyzed our results, linking human attention (obtained through eye-tracking) with XAI saliency on CV models and brain region activations. We show how a parallel analysis of human and computer vision can provide useful information for both the neuroscience community (allocation theory) and the ML community (biological plausibility of convolutional models).

Regulators, researchers, and practitioners recognize the urgency of explainability in artificial intelligence systems, including the ones based on machine learning for graph-structured data. Despite the large number of proposals, however, a common understanding of what constitutes a good explanation is still lacking: different explainers often arrive at different conclusions on the same problem instance, making it hard for practitioners to choose among them. Furthermore, explainers often produce explanations through opaque logic hard to understand and assess -- ironically mirroring the black box nature they aim to elucidate. Recent proposals in the literature for benchmarking graph-based explainers typically involve embedding specific logic into data, training a black-box model, and then empirically assessing how well the explanation matches the embedded logic, i.e., they test truthfulness to the data. In contrast, we propose a true-to-the-model axiomatic framework for auditing explainers in the task of node classification on graphs. Our proposal hinges on the fundamental idea that an explainer should discern if a model relies on a particular feature for classifying a node. Building on this concept, we develop three types of white-box classifiers, with clear internal logic, that are relevant in real-world applications. We then formally prove that the set of features that can induce a change in the classification correctly corresponds to a ground-truth set of predefined important features. This property allows us to use the white-box classifiers to build a testing framework. We apply this framework to both synthetic and real data and evaluate various state-of-the-art explainers, thus characterizing their behavior. Our findings highlight how explainers often react in a rather counter-intuitive fashion to technical details that might be easily overlooked. Our approach offers valuable insights and recommended practices for selecting the right explainer given the task at hand, and for developing new methods for explaining graph-learning models.

Cycling is an outdoor activity with massive health benefits, and an effective solution for sustainable urban transport. Despite these benefits and the recent rising popularity of cycling, most countries still have a negligible uptake. This uptake is especially low for women: there is a largely unexplained, persistent gender gap in cycling. To understand the determinants of this gender gap in cycling at scale, here we use massive, automatically-collected data from the tracking application Strava on outdoor cycling for 61 cities across the United States, the United Kingdom, Italy and the Benelux area. While Strava data is particularly well-suited to describe the behavior of regular cyclists and its generalizability to occasional cyclists requires further investigation, the size of these data and their characteristics represent an unprecedented opportunity for the literature on cycling. Leveraging the associated gender and usage information, we first quantify the emerging gender gap in recreational cycling at city-level. A comparison of cycling rates of women across cities within similar geographical areas—where the penetration of Strava is assumed to be comparable—unveils a broad range of gender gaps. On a macroscopic level, we link this heterogeneity to a variety of urban indicators and provide evidence for traditional hypotheses on the determinants of the gender-cycling-gap. We find a positive association between female cycling rate and urban road safety. On a microscopic level, we identify female preferences for street-specific features in the city of New York. Assuming that the determinants of the gender-cycling-gap are similar across regular and occasional cyclists, our study suggests that enhancing the quality of the dedicated cycling infrastructure may be a way to make urban environments more accessible for women, thereby making urban transport more sustainable for everyone.

Convolutional Neural Networks (CNN) are a class of machine learning models predominately used in computer vision tasks and can achieve human-like performance through learning from experience. Their striking similarities to the structural and functional principles of the primate visual system allow for comparisons between these artificial networks and their biological counterparts, enabling exploration of how visual functions and neural representations may emerge in the real brain from a limited set of computational principles. After considering the basic features of CNNs, we discuss the opportunities and challenges of endorsing CNNs as in silico models of the primate visual system. Specifically, we highlight several emerging notions about the anatomical and physiological properties of the visual system that still need to be systematically integrated into current CNN models. These tenets include the implementation of parallel processing pathways from the early stages of retinal input and the reconsideration of several assumptions concerning the serial progression of information flow. We suggest design choices and architectural constraints that could facilitate a closer alignment with biology provide causal evidence of the predictive link between the artificial and biological visual systems. Adopting this principled perspective could potentially lead to new research questions and applications of CNNs beyond modeling object recognition.

Convolutional Neural Networks (CNNs) are nowadays the model of choice in Computer Vision, thanks to their ability to automatize the feature extraction process in visual tasks. However, the knowledge acquired during training is fully sub-symbolic, and hence difficult to understand and explain to end users. In this paper, we propose a new technique called HOLMES (HOLonym-MEronym based Semantic inspection) that decomposes a label into a set of related concepts, and provides component-level explanations for an image classification model. Specifically, HOLMES leverages ontologies, web scraping and transfer learning to automatically construct meronym (parts)-based detectors for a given holonym (class). Then, it produces heatmaps at the meronym level and finally, by probing the holonym CNN with occluded images, it highlights the importance of each part on the classification output. Compared to state-of-the-art saliency methods, HOLMES takes a step further and provides information about both where and what the holonym CNN is looking at. It achieves so without relying on densely annotated datasets and without forcing concepts to be associated to single computational units. Extensive experimental evaluation on different categories of objects (animals, tools and vehicles) shows the feasibility of our approach. On average, HOLMES explanations include at least two meronyms, and the ablation of a single meronym roughly halves the holonym model confidence. The resulting heatmaps were quantitatively evaluated using the deletion/insertion/preservation curves. All metrics were comparable to those achieved by GradCAM, while offering the advantage of further decomposing the heatmap in human-understandable concepts. In addition, results were largely above chance level, thus highlighting both the relevance of meronyms to object classification, as well as HOLMES ability to capture it. The code is available at https://github.com/FrancesC0de/HOLMES .

eXplainable AI (XAI) involves two intertwined but separate challenges: the development of techniques to extract explanations from black-box AI models and the way such explanations are presented to users, i.e., the explanation user interface. Despite its importance, the second aspect has received limited attention so far in the literature. Effective AI explanation interfaces are fundamental for allowing human decision-makers to take advantage and oversee high-risk AI systems effectively. Following an iterative design approach, we present the first cycle of prototyping-testing-redesigning of an explainable AI technique and its explanation user interface for clinical Decision Support Systems (DSS). We first present an XAI technique that meets the technical requirements of the healthcare domain: sequential, ontology-linked patient data, and multi-label classification tasks. We demonstrate its applicability to explain a clinical DSS, and we design a first prototype of an explanation user interface. Next, we test such a prototype with healthcare providers and collect their feedback with a two-fold outcome: First, we obtain evidence that explanations increase users’ trust in the XAI system, and second, we obtain useful insights on the perceived deficiencies of their interaction with the system, so we can re-design a better, more human-centered explanation interface.

The PRE-ACT project is a newly launched Horizon Europe project that aims to use Artificial Intelligence (AI) towards predicting the risk of side effects of radiotherapy treatment for breast cancer patients. In this paper, we outline four main threads pertaining to AI and computing that are part of the project's research agenda, namely: (i) Explainable AI techniques to make the risk prediction interpretable for the patient and the clinician; (ii) Fair AI techniques to identify and explain potential biases in clinical decision support systems; (iii) Training of AI models from distributed data through Federated Learning algorithms to ensure data privacy; (iv) Mobile applications to provide the patients and clinicians with an interface for the side effect risk prediction. For each of these directions, we provide an overview of the state-of-the-art, with emphasis on techniques that are more relevant for the project. Collectively, these four threads can be seen as enforcing Trustworthy AI and pave the way to transparent and responsible AI systems that are adopted by end-users and may thus unleash the full potential of AI.

Most methods for explaining black-box classifiers (e.g., on tabular data, images, or time series) rely on measuring the impact that removing/perturbing features has on the model output. This forces the explanation language to match the classifier's feature space. However, when dealing with graph data, in which the basic features correspond to the edges describing the graph structure, this matching between features space and explanation language might not be appropriate. Decoupling the feature space (edges) from a desired high-lever explanation language (such as motifs) is thus a major challenge towards developing actionable explanations for graph classification tasks. In this paper we introduce Graphshap, a Shapley-based approach able to provide motif-based explanations for identityaware graph classifiers, assuming no knowledge whatsoever about the model or its training data: the only requirement is that the classifier can be queried as a black-box at will. For the sake of computational efficiency we explore a progressive approximation strategy and show how a simple kernel can efficiently approximate explanation scores, thus allowing Graphshap to scale on scenarios with a large explanation space (i.e., large number of motifs). We showcase Graphshap on a real-world brain-network dataset consisting of patients affected by Autism Spectrum Disorder and a control group. Our experiments highlight how the classification provided by a black-box model can be effectively explained by few connectomics patterns.

Images are loaded with semantic information that pertains to real-world ontologies: dog breeds share mammalian similarities, food pictures are often depicted in domestic environments, and so on. However, when training machine learning models for image classification, the relative similarities amongst object classes are commonly paired with one-hot-encoded labels. According to this logic, if an image is labelled as spoon, then tea-spoon and shark are equally wrong in terms of training loss. To overcome this limitation, we explore the integration of additional goals that reflect ontological and semantic knowledge, improving model interpretability and trustworthiness. We suggest a generic approach that allows to derive an additional loss term starting from any kind of semantic information about the classification label. First, we show how to apply our approach to ontologies and word embeddings, and discuss how the resulting information can drive a supervised learning process. Second, we use our semantically enriched loss to train image classifiers, and analyse the trade-offs between accuracy, mistake severity, and learned internal representations. Finally, we discuss how this approach can be further exploited in terms of explainability and adversarial robustness.