Author name cluster

Alberto Termine

Possible papers associated with this exact author name in Arrow. This page groups case-insensitive exact name matches and is not a full identity disambiguation profile.

4 papers

2 author rows

ICLR Conference 2025 Conference Paper

Causal Concept Graph Models: Beyond Causal Opacity in Deep Learning

Gabriele Dominici
Pietro Barbiero
Mateo Espinosa Zarlenga
Alberto Termine
Martin Gjoreski
Giuseppe Marra
Marc Langheinrich

Causal opacity denotes the difficulty in understanding the "hidden" causal structure underlying the decisions of deep neural network (DNN) models. This leads to the inability to rely on and verify state-of-the-art DNN-based systems, especially in high-stakes scenarios. For this reason, circumventing causal opacity in DNNs represents a key open challenge at the intersection of deep learning, interpretability, and causality. This work addresses this gap by introducing Causal Concept Graph Models (Causal CGMs), a class of interpretable models whose decision-making process is causally transparent by design. Our experiments show that Causal CGMs can: (i) match the generalisation performance of causally opaque models, (ii) enable human-in-the-loop corrections to mispredicted intermediate reasoning steps, boosting not just downstream accuracy after corrections but also the reliability of the explanations provided for specific instances, and (iii) support the analysis of interventional and counterfactual scenarios, thereby improving the model's causal interpretability and supporting the effective verification of its reliability and fairness.

Details

NeurIPS Conference 2025 Conference Paper

Causally Reliable Concept Bottleneck Models

Giovanni De Felice
Arianna Casanova Flores
Francesco De Santis
Silvia Santini
Johannes Schneider
Pietro Barbiero
Alberto Termine

Concept-based models are an emerging paradigm in deep learning that constrains the inference process to operate through human-interpretable variables, facilitating explainability and human interaction. However, these architectures, on par with popular opaque neural models, fail to account for the true causal mechanisms underlying the target phenomena represented in the data. This hampers their ability to support causal reasoning tasks, limits out-of-distribution generalization, and hinders the implementation of fairness constraints. To overcome these issues, we propose Causally reliable Concept Bottleneck Models (C$^2$BMs), a class of concept-based architectures that enforce reasoning through a bottleneck of concepts structured according to a model of the real-world causal mechanisms. We also introduce a pipeline to automatically learn this structure from observational data and unstructured background knowledge (e. g. , scientific literature). Experimental evidence suggests that C$^2$BMs are more interpretable, causally reliable, and improve responsiveness to interventions w. r. t. standard opaque and concept-based models, while maintaining their accuracy.

PDF Details

EUMAS Conference 2021 Conference Paper

Logic and Model Checking by Imprecise Probabilistic Interpreted Systems

Alberto Termine
Alessandro Antonucci 0001
Giuseppe Primiero
Alessandro Facchini

Abstract Stochastic multi-agent systems raise the necessity to extend probabilistic model checking to the epistemic domain. Results in this direction have been achieved by epistemic extensions of Probabilistic Computation Tree Logic and related Probabilistic Interpreted Systems. The latter, however, suffer of an important limitation: they require the probabilities governing the system’s behaviour to be fully specified. A promising way to overcome this limitation is represented by imprecise probabilities. In this paper we introduce imprecise probabilistic interpreted systems and present a related logical language and model-checking procedures based on recent advances in the study of imprecise Markov processes.

Details

LORI Conference 2021 Conference Paper

Modelling Accuracy and Trustworthiness of Explaining Agents

Alberto Termine
Giuseppe Primiero
Fabio Aurelio D'Asaro

Abstract Current research in Explainable AI includes post-hoc explanation methods that focus on building transparent explaining agents able to emulate opaque ones. Such agents are naturally required to be accurate and trustworthy. However, what it means for an explaining agent to be accurate and trustworthy is far from being clear. We characterize accuracy and trustworthiness as measures of the distance between the formal properties of a given opaque system and those of its transparent explanantes. To this aim, we extend Probabilistic Computation Tree Logic with operators to specify degrees of accuracy and trustworthiness of explaining agents. We also provide a semantics for this logic, based on a multi-agent structure and relative model-checking algorithms. The paper concludes with a simple example of a possible application.

Details