arXiv subjects

Roberto Pellungrini

Publications and source records attributed to Roberto Pellungrini.

10 recordsLinked to original sources

arXiv2026

Comparing Explanations is Not Enough, Explain the Change: New Standards are Needed to Explain Behavioral Shifts in Large Language Models

Large-scale foundation models exhibit behavioral shifts when subjected to interventions such as scaling, fine-tuning, reinforcement learning with human feedback, or in-context learning. Current explainability methods are structurally ill-suited to explain these shifts, because they either treat models as static objects, as traditional eXplainable AI (XAI) approaches do, or merely compare independent explanations across different checkpoints of a model. As a result, these approaches fail to explain the functional transition between two model instances in which a certain behavior has shifted following an intervention. This gap creates significant governance risks across jurisdictions including the EU AI Act, US state legislation, and Chinese AI regulations, which require documenting causal chains for substantial system modifications. This position paper argues that explaining behavioral shifts in large language models requires a principled approach that treats the shift itself as the primary object of explanation: namely, one that explains how and why an intervention transforms a reference model into an updated model with different behavior. To support this claim, we introduce Comparative XAI (XAI$_Δ$), a novel XAI paradigm aimed at explaining the difference between two model checkpoints where a behavior has shifted, together with a set of desiderata specifying what XAI$_Δ$ explainers and explanations must satisfy, including comparability, validity, actionability, and monitoring, with the goal of grounding model auditing in explicit, measurable requirements. Finally, we provide preliminary evidence suggesting the need for XAI$_Δ$ in practice through illustrative experiments, compiling the resulting findings into a transition report directly usable for governance and incident documentation.

cs.AI↗

arXiv2025

To Ask or Not to Ask: Learning to Require Human Feedback

Developing decision-support systems that complement human performance in classification tasks remains an open challenge. A popular approach, Learning to Defer (LtD), allows a Machine Learning (ML) model to pass difficult cases to a human expert. However, LtD treats humans and ML models as mutually exclusive decision-makers, restricting the expert contribution to mere predictions. To address this limitation, we propose Learning to Ask (LtA), a new framework that handles both when and how to incorporate expert input in an ML model. LtA is based on a two-part architecture: a standard ML model and an enriched model trained with additional expert human feedback, with a formally optimal strategy for selecting when to query the enriched model. We provide two practical implementations of LtA: a sequential approach, which trains the models in stages, and a joint approach, which optimises them simultaneously. For the latter, we design surrogate losses with realisable-consistency guarantees. Our experiments with synthetic and real expert data demonstrate that LtA provides a more flexible and powerful foundation for effective human-AI collaboration.

cs.LG↗

arXiv2025

One-Shot Clustering for Federated Learning Under Clustering-Agnostic Assumption

Federated Learning (FL) is a widespread and well-adopted paradigm of decentralised learning that allows training one model from multiple sources without the need to transfer data between participating clients directly. Since its inception in 2015, it has been divided into numerous subfields that deal with application-specific issues, such as data heterogeneity or resource allocation. One such sub-field, Clustered Federated Learning (CFL), deals with the problem of clustering the population of clients into separate cohorts to deliver personalised models. Although a few remarkable works have been published in this domain, the problem remains largely unexplored, as its basic assumptions and settings differ slightly from those of standard FL. In this work, we present One-Shot Clustered Federated Learning (OCFL), a clustering-agnostic algorithm that can automatically detect the earliest suitable moment for clustering. Our algorithm is based on computing the cosine distance between the gradients of the clients and a temperature measure that detects when the federated model starts to converge. We empirically evaluate our methodology by testing various one-shot clustering algorithms for over forty different tasks on five benchmark datasets. Our experiments showcase the good performance of our approach when used to perform CFL in an automated manner without the need to adjust hyperparameters. We also revisit the practical feasibility of CFL algorithms based on the gradients of the clients, providing firm evidence of the high efficiency of density-based clustering methods when used to differentiate between the loss surfaces of neural networks trained on different distributions. Moreover, by inspecting the feasibility of local explanations generated with the help of GradCAM, we can provide more insights into the relationship between personalisation and the explainability of local predictions.

cs.LG↗

arXiv2025

One-Shot Clustering for Federated Learning

Federated Learning (FL) is a widespread and well adopted paradigm of decentralized learning that allows training one model from multiple sources without the need to directly transfer data between participating clients. Since its inception in 2015, it has been divided into numerous sub-fields that deal with application-specific issues, be it data heterogeneity or resource allocation. One such sub-field, Clustered Federated Learning (CFL), is dealing with the problem of clustering the population of clients into separate cohorts to deliver personalized models. Although few remarkable works have been published in this domain, the problem is still largely unexplored, as its basic assumption and settings are slightly different from standard FL. In this work, we present One-Shot Clustered Federated Learning (OCFL), a clustering-agnostic algorithm that can automatically detect the earliest suitable moment for clustering. Our algorithm is based on the computation of cosine similarity between gradients of the clients and a temperature measure that detects when the federated model starts to converge. We empirically evaluate our methodology by testing various one-shot clustering algorithms for over thirty different tasks on three benchmark datasets. Our experiments showcase the good performance of our approach when used to perform CFL in an automated manner without the need to adjust hyperparameters.

cs.LG↗

arXiv2025

Interpretable and Fair Mechanisms for Abstaining Classifiers

Abstaining classifiers have the option to refrain from providing a prediction for instances that are difficult to classify. The abstention mechanism is designed to trade off the classifier's performance on the accepted data while ensuring a minimum number of predictions. In this setting, often fairness concerns arise when the abstention mechanism solely reduces errors for the majority groups of the data, resulting in increased performance differences across demographic groups. While there exist a bunch of methods that aim to reduce discrimination when abstaining, there is no mechanism that can do so in an explainable way. In this paper, we fill this gap by introducing Interpretable and Fair Abstaining Classifier IFAC, an algorithm that can reject predictions both based on their uncertainty and their unfairness. By rejecting possibly unfair predictions, our method reduces error and positive decision rate differences across demographic groups of the non-rejected data. Since the unfairness-based rejections are based on an interpretable-by-design method, i.e., rule-based fairness checks and situation testing, we create a transparent process that can empower human decision-makers to review the unfair predictions and make more just decisions for them. This explainable aspect is especially important in light of recent AI regulations, mandating that any high-risk decision task should be overseen by human experts to reduce discrimination risks.

cs.LG↗

arXiv2025

Deferring Concept Bottleneck Models: Learning to Defer Interventions to Inaccurate Experts

Concept Bottleneck Models (CBMs) are machine learning models that improve interpretability by grounding their predictions on human-understandable concepts, allowing for targeted interventions in their decision-making process. However, when intervened on, CBMs assume the availability of humans that can identify the need to intervene and always provide correct interventions. Both assumptions are unrealistic and impractical, considering labor costs and human error-proneness. In contrast, Learning to Defer (L2D) extends supervised learning by allowing machine learning models to identify cases where a human is more likely to be correct than the model, thus leading to deferring systems with improved performance. In this work, we gain inspiration from L2D and propose Deferring CBMs (DCBMs), a novel framework that allows CBMs to learn when an intervention is needed. To this end, we model DCBMs as a composition of deferring systems and derive a consistent L2D loss to train them. Moreover, by relying on a CBM architecture, DCBMs can explain why defer occurs on the final task. Our results show that DCBMs achieve high predictive performance and interpretability at the cost of deferring more to humans.

cs.LG↗

arXiv2025

AI, Meet Human: Learning Paradigms for Hybrid Decision Making Systems

Everyday we increasingly rely on machine learning models to automate and support high-stake tasks and decisions. This growing presence means that humans are now constantly interacting with machine learning-based systems, training and using models everyday. Several different techniques in computer science literature account for the human interaction with machine learning systems, but their classification is sparse and the goals varied. This survey proposes a taxonomy of Hybrid Decision Making Systems, providing both a conceptual and technical framework for understanding how current computer science literature models interaction between humans and machines.

cs.LG↗

arXiv2025

Mathematical Foundation of Interpretable Equivariant Surrogate Models

This paper introduces a rigorous mathematical framework for neural network explainability, and more broadly for the explainability of equivariant operators called Group Equivariant Operators (GEOs) based on Group Equivariant Non-Expansive Operators (GENEOs) transformations. The central concept involves quantifying the distance between GEOs by measuring the non-commutativity of specific diagrams. Additionally, the paper proposes a definition of interpretability of GEOs according to a complexity measure that can be defined according to each user preferences. Moreover, we explore the formal properties of this framework and show how it can be applied in classical machine learning scenarios, like image classification with convolutional neural networks.

stat.ML↗

arXiv2024

Bridging the Gap in Hybrid Decision-Making Systems

We introduce BRIDGET, a novel human-in-the-loop system for hybrid decision-making, aiding the user to label records from an un-labeled dataset, attempting to ``bridge the gap'' between the two most popular Hybrid Decision-Making paradigms: those featuring the human in a leading position, and the other with a machine making most of the decisions. BRIDGET understands when either a machine or a human user should be in charge, dynamically switching between two statuses. In the different statuses, BRIDGET still fosters the human-AI interaction, either having a machine learning model assuming skeptical stances towards the user and offering them suggestions, or towards itself and calling the user back. We believe our proposal lays the groundwork for future synergistic systems involving a human and a machine decision-makers.

cs.AI↗

arXiv2021

Scikit-mobility: a Python library for the analysis, generation and risk assessment of mobility data

The last decade has witnessed the emergence of massive mobility data sets, such as tracks generated by GPS devices, call detail records, and geo-tagged posts from social media platforms. These data sets have fostered a vast scientific production on various applications of mobility analysis, ranging from computational epidemiology to urban planning and transportation engineering. A strand of literature addresses data cleaning issues related to raw spatiotemporal trajectories, while the second line of research focuses on discovering the statistical "laws" that govern human movements. A significant effort has also been put on designing algorithms to generate synthetic trajectories able to reproduce, realistically, the laws of human mobility. Last but not least, a line of research addresses the crucial problem of privacy, proposing techniques to perform the re-identification of individuals in a database. A view on state of the art cannot avoid noticing that there is no statistical software that can support scientists and practitioners with all the aspects mentioned above of mobility data analysis. In this paper, we propose scikit-mobility, a Python library that has the ambition of providing an environment to reproduce existing research, analyze mobility data, and simulate human mobility habits. scikit-mobility is efficient and easy to use as it extends pandas, a popular Python library for data analysis. Moreover, scikit-mobility provides the user with many functionalities, from visualizing trajectories to generating synthetic data, from analyzing statistical patterns to assessing the privacy risk related to the analysis of mobility data sets.

physics.soc-ph↗