Decision trees are a popular machine learning method, valued for their inherent explainability. In Explainable AI, decision trees serve as surrogate models for complex black box AI models or as approximations of parts of such models. A key challenge of this approach is assessing how accurately the extracted decision tree represents the original model and determining the extent to which it can be trusted as an approximation of its behaviour. In this work, we investigate the use of the Probably Approximately Correct (PAC) framework to provide a theoretical guarantee of fidelity for decision trees extracted from AI models. Leveraging the theoretical foundations of the PAC framework, we adapt a decision tree algorithm to ensure a PAC guarantee under specific conditions. We focus on binary classification and conduct experiments where we extract decision trees from BERT-based language models with PAC guarantees. Our results indicate occupational gender bias in these models, which confirm previous results in the literature. Additionally, the decision tree format enhances the visualization of which occupations are most impacted by social bias.

Deep neural networks are widely used in practical applications of AI, however, their inner structure and complexity made them generally not easily interpretable. Model transparency and interpretability are key requirements for multiple scenarios where high performance is not enough to adopt the proposed solution. In this work, a differentiable approximation of L0regularization is adapted into a logic-based neural network, the Multi-layer Logical Perceptron (MLLP), to study its efficacy in reducing the complexity of its discrete interpretable version, the Concept Rule Set (CRS), while retaining its performance. The results are compared to alternative heuristics like Random Binarization of the network weights, to determine if better results can be achieved when using a less-noisy technique that sparsifies the network based on the loss function instead of a random distribution. The trade-off between the CRS complexity and its performance is discussed.

Deep vision models often rely on biases learned from spurious correlations in datasets. To identify these biases, methods that interpret high-level, human-understandable concepts are more effective than those relying primarily on low-level features like heatmaps. A major challenge for these concept-based methods is the lack of image annotations indicating potentially bias-inducing concepts, since creating such annotations requires detailed labeling for each dataset and concept, which is highly labor-intensive. We present CUBIC (Concept embeddings for Unsupervised Bias IdentifiCation), a novel method that automatically discovers interpretable concepts that may bias classifier behavior. Unlike existing approaches, CUBIC does not rely on predefined bias candidates or examples of model failures tied to specific biases, as these are not always available in the data. Instead, it utilizes image-text latent space and linear classifier probes to examine how the latent representation of a superclass label—shared by all instances in the dataset—is influenced by the presence of a concept. By measuring these shifts against the normal vector to the classifier’s decision boundary, CUBIC identifies concepts that significantly influence model predictions. Our experiments demonstrate that CUBIC effectively uncovers previously unknown biases using Vision-Language Models (VLMs) without requiring the samples in the dataset where the classifier underperforms or prior knowledge of potential biases.

Post-hoc methods in Explainable AI (XAI) elucidate black-box models by identifying input features critical to the model's decision-making. Recent advancements in these methods have facilitated the generation of logic-based explanations that capture interactions among input features. However, these techniques often encounter critical limitations, notably the inability to ensure logical consistency and fidelity between generated explanations and the model's actual decision-making processes. Such inconsistencies jeopardize the reliability of explanations particularly in high-risk domains. To address this gap, we introduce a novel, theoretically rigorous approach rooted in category theory. Specifically, we propose the concept of an explaining functor, which preserves logical entailment structurally between the explanations and the decisions of black-box models. By establishing a categorical framework, our method guarantees the coherence and accuracy of extracted explanations, thus overcoming the common pitfalls associated with heuristic-based explanation methods. We demonstrate the practical efficacy of our theoretical contributions through two synthetic benchmarks that highlight significant reductions in contradictory and unfaithful explanations. Our experiments show how our framework can provide mathematically grounded, compositional, and coherent explanations.

We introduce a novel framework to deal with fairness, accountability and explainability of intelligent systems. This framework puts together several tools to deal with bias at the level of data, algorithms and human cognition. The framework makes use of intelligent classifiers endowed with fuzzy-grounded linguistic explainability. As a result, it facilitates the exhaustive comparison of (white/grey/black)-box modelling techniques in combination with different strategies for handling missing values and unbalanced datasets. The proposal is evaluated on a realworld dataset in the context of banking services and reported results are encouraging.

The interest in explainable artificial intelligence has grown strongly in recent years because of the need to convey safety and trust in the ‘how’ and ‘why’ of automated decision-making to users. While a plethora of approaches has been developed, only a few focus on how to use domain knowledge and how this influences the understanding of explanations by users. In this paper, we show that by using ontologies we can improve the human understandability of global post-hoc explanations, presented in the form of decision trees. In particular, we introduce Trepan Reloaded, which builds on Trepan, an algorithm that extracts surrogate decision trees from black-box models. Trepan Reloaded includes ontologies, that model domain knowledge, in the process of extracting explanations to improve their understandability. We tested the understandability of the extracted explanations by humans in a user study with four different tasks. We evaluate the results in terms of response times and correctness, subjective ease of understanding and confidence, and similarity of free text responses. The results show that decision trees generated with Trepan Reloaded, taking into account domain knowledge, are significantly more understandable throughout than those generated by standard Trepan. The enhanced understandability of post-hoc explanations is achieved with little compromise on the accuracy with which the surrogate decision trees replicate the behaviour of the original neural network models.

Explainability in Artificial Intelligence has been revived as a topic of active research by the need of conveying safety and trust to users in the `how' and `why' of automated decision-making. Whilst a plethora of approaches have been developed for post-hoc explainability, only a few focus on how to use domain knowledge, and how this influences the understandability of global explanations from the users' perspective. In this paper, we show how ontologies help the understandability of global post-hoc explanations, presented in the form of symbolic models. In particular, we build on Trepan, an algorithm that explains artificial neural networks by means of decision trees, and we extend it to include ontologies modeling domain knowledge in the process of generating explanations. We present the results of a user study that measures the understandability of decision trees using a syntactic complexity measure, and through time and accuracy of responses as well as reported user confidence and understandability. The user study considers domains where explanations are critical, namely, in finance and medicine. The results show that decision trees generated with our algorithm, taking into account domain knowledge, are more understandable than those generated by standard Trepan without the use of ontologies.

Ontology engineering is a hard and error-prone task, in which small changes may lead to errors, or even produce an inconsistent ontology. As ontologies grow in size, the need for automated methods for repairing inconsistencies while preserving as much of the original knowledge as possible increases. Most previous approaches to this task are based on removing a few axioms from the ontology to regain consistency. We propose a new method based on weakening these axioms to make them less restrictive, employing the use of refinement operators. We introduce the theoretical framework for weakening DL ontologies, propose algorithms to repair ontologies based on the framework, and provide an analysis of the computational complexity. Through an empirical analysis made over real-life ontologies, we show that our approach preserves significantly more of the original knowledge of the ontology than removing axioms.

Axiom weakening is a novel technique that allows for fine-grained repair of inconsistent ontologies. In a multi-agent setting, integrating ontologies corresponding to multiple agents may lead to inconsistencies. Such inconsistencies can be resolved after the integrated ontology has been built, or their generation can be prevented during ontology generation. We implement and compare these two approaches. First, we study how to repair an inconsistent ontology resulting from a voting-based aggregation of views of heterogeneous agents. Second, we prevent the generation of inconsistencies by letting the agents engage in a turn-based rational protocol about the axioms to be added to the integrated ontology. We instantiate the two approaches using real-world ontologies and compare them by measuring the levels of satisfaction of the agents w.r.t. the ontology obtained by the two procedures.

We present a computational framework for conceptual blending , a concept invention method that is advocated in cognitive science as a fundamental and uniquely human engine for creative thinking. Our framework treats a crucial part of the blending process, namely the generalisation of input concepts, as a search problem that is solved by means of modern answer set programming methods to find commonalities among input concepts. We also address the problem of pruning the space of possible blends by introducing metrics that capture most of the so-called optimality principles , described in the cognitive science literature as guidelines to produce meaningful and serendipitous blends. As a proof of concept, we demonstrate how our system invents novel concepts and theories in domains where creativity is crucial, namely mathematics and music.

This paper proposes the notion of experience to help situate agents in their environment, providing a link on how the continually evolving environment impacts the evolution of an agent's BDI model and vice versa. Then, using the notion of shared experience as a primitive construct, we develop a novel formal model of shared intention which we believe more adequately describes social behaviour than traditional BDI logics that focus on individual agents. Whilst many philosophers have argued that collective intentionality cannot always be equated to the collection of the individual agents' intentions, there has been no AI model that addresses this issue. We believe this is the first attempt to develop an explicit notion of shared experience from an AI perspective.

In this paper, we define a class of nested logic programs, called Nested Logic Programs with Ordered Disjunction (LPODs + ), which makes it possible to specify conditional (qualitative) preferences by means of nested preference statements. To this end, we augment the syntax of Logic Programs with Ordered Disjunction (LPODs) to capture more general expressions. We define the LPODs + semantics in a simple way and we extend most of the results of LPODs showing how our approach generalizes the LPODs framework in a proper way. We also show how the LPODs + semantics can be computed in terms of a translation procedure that maps a nested ordered disjunction program (OD + -program) into a disjunctive logic program.

In this paper, we introduce a possibilistic argumentation-based decision making framework which is able to capture uncertain information and exceptions/defaults. In particular, we define the concept of a possibilistic decision making framework which is based on a possibilistic default theory, a set of decisions and a set of prioritized goals. This set of goals captures user preferences related to the achievement of a particular state in a decision making problem. By considering the inference of the possibilistic well-founded semantics, the concept of argument with respect to a decision is defined. This argument captures the feasibility of reaching a goal by applying a decision in a given context. The inference in the argumentation decision making framework is based on basic argumentation semantics. Since some basic argumentation semantics can infer more than one possible scenario of a possibilistic decision making problem, we define some criteria for selecting potential solutions of the problem.