Marco Maggini

ECAI Conference 2025 Conference Paper

Pirates of the RAG: Adaptively Attacking LLMs to Leak Knowledge Bases

• Christian Di Maio
• Cristian Cosci
• Marco Maggini
• Valentina Poggioni
• Stefano Melacci

The growing ubiquity of Retrieval-Augmented Generation (RAG) systems in several real-world services triggers severe concerns about their security. A RAG system improves the generative capabilities of a Large Language Model (LLM) by a retrieval mechanism that operates on a private knowledge base, whose unintended exposure could lead to severe consequences, including breaches of private and sensitive information. This paper presents a black-box attack to force a RAG system to leak its private knowledge base which, unlike existing approaches, is both adaptive and automatic. A relevance-based mechanism and an attacker-side open-source LLM favor the generation of effective queries to leak most of the (hidden) knowledge base. Extensive experimentation proves the quality of the proposed algorithm in different RAG pipelines and domains, compared to very recent related approaches, which turn out to be either not fully black-box, not adaptive, or not based on open-source models. The findings from our study highlight the urgent need for more robust privacy safeguards in the design and deployment of RAG systems. We have made the open-source code for our experimental procedure available for public use [12].

NeSy Conference 2023 Conference Paper

Logic Explained Networks

• Gabriele Ciravegna
• Pietro Barbiero
• Francesco Giannini
• Marco Gori
• Pietro Liò
• Marco Maggini
• Stefano Melacci

The large and still increasing popularity of deep learning clashes with a major limit of neural network architectures, that consists in their lack of capability in providing human-understandable motivations of their decisions. In situations in which the machine is expected to support the decision of human experts, providing a comprehensible explanation is a feature of crucial importance. The language used to communicate the explanations must be formal enough to be implementable in a machine and friendly enough to be understandable by a wide audience. In this paper, we propose a general approach to Explainable Artificial Intelligence in the case of neural architectures, showing how a mindful design of the networks leads to a family of interpretable deep learning models called Logic Explained Networks (LENs). LENs only require their inputs to be human-understandable predicates, and they provide explanations in terms of simple First-Order Logic (FOL) formulas involving such predicates. LENs are general enough to cover a large number of scenarios. Amongst them, we consider the case in which LENs are directly used as special classifiers with the capability of being explainable, or when they act as additional networks with the role of creating the conditions for making a black-box classifier explainable by FOL formulas. Despite supervised learning problems are mostly emphasized, we also show that LENs can learn and provide explanations in unsupervised learning settings. Experimental results on several datasets and tasks show that LENs may yield better classifications than established white-box models, such as decision trees and Bayesian rule lists, while providing more compact and meaningful explanations.

AIJ Journal 2023 Journal Article

Logic Explained Networks

• Gabriele Ciravegna
• Pietro Barbiero
• Francesco Giannini
• Marco Gori
• Pietro Liò
• Marco Maggini
• Stefano Melacci

The large and still increasing popularity of deep learning clashes with a major limit of neural network architectures, that consists in their lack of capability in providing human-understandable motivations of their decisions. In situations in which the machine is expected to support the decision of human experts, providing a comprehensible explanation is a feature of crucial importance. The language used to communicate the explanations must be formal enough to be implementable in a machine and friendly enough to be understandable by a wide audience. In this paper, we propose a general approach to Explainable Artificial Intelligence in the case of neural architectures, showing how a mindful design of the networks leads to a family of interpretable deep learning models called Logic Explained Networks (LENs). LENs only require their inputs to be human-understandable predicates, and they provide explanations in terms of simple First-Order Logic (FOL) formulas involving such predicates. LENs are general enough to cover a large number of scenarios. Amongst them, we consider the case in which LENs are directly used as special classifiers with the capability of being explainable, or when they act as additional networks with the role of creating the conditions for making a black-box classifier explainable by FOL formulas. Despite supervised learning problems are mostly emphasized, we also show that LENs can learn and provide explanations in unsupervised learning settings. Experimental results on several datasets and tasks show that LENs may yield better classifications than established white-box models, such as decision trees and Bayesian rule lists, while providing more compact and meaningful explanations.

AAAI Conference 2020 Conference Paper

A Constraint-Based Approach to Learning and Explanation

• Gabriele Ciravegna
• Francesco Giannini
• Stefano Melacci
• Marco Maggini
• Marco Gori

In the last few years we have seen a remarkable progress from the cultivation of the idea of expressing domain knowledge by the mathematical notion of constraint. However, the progress has mostly involved the process of providing consistent solutions with a given set of constraints, whereas learning “new” constraints, that express new knowledge, is still an open challenge. In this paper we propose a novel approach to learning of constraints which is based on information theoretic principles. The basic idea consists in maximizing the transfer of information between task functions and a set of learnable constraints, implemented using neural networks subject to L1 regularization. This process leads to the unsupervised development of new constraints that are fulfilled in different subportions of the input domain. In addition, we define a simple procedure that can explain the behaviour of the newly devised constraints in terms of First-Order Logic formulas, thus extracting novel knowledge on the relationships between the original tasks. An experimental evaluation is provided to support the proposed approach, in which we also explore the regularization effects introduced by the proposed Information- Based Learning of Constraint (IBLC) algorithm.

ECAI Conference 2020 Conference Paper

A Lagrangian Approach to Information Propagation in Graph Neural Networks

• Matteo Tiezzi
• Giuseppe Marra
• Stefano Melacci
• Marco Maggini
• Marco Gori

In many real world applications, data are characterized by a complex structure, that can be naturally encoded as a graph. In the last years, the popularity of deep learning techniques has renewed the interest in neural models able to process complex patterns. In particular, inspired by the Graph Neural Network (GNN) model, different architectures have been proposed to extend the original GNN scheme. GNNs exploit a set of state variables, each assigned to a graph node, and a diffusion mechanism of the states among neighbor nodes, to implement an iterative procedure to compute the fixed point of the (learnable) state transition function. In this paper, we propose a novel approach to the state computation and the learning algorithm for GNNs, based on a constraint optimisation task solved in the Lagrangian framework. The state convergence procedure is implicitly expressed by the constraint satisfaction mechanism and does not require a separate iterative phase for each epoch of the learning procedure. In fact, the computational structure is based on the search for saddle points of the Lagrangian in the adjoint space composed of weights, neural outputs (node states), and Lagrange multipliers. The proposed approach is compared experimentally with other popular models for processing graphs.

NeurIPS Conference 2020 Conference Paper

Focus of Attention Improves Information Transfer in Visual Features

• Matteo Tiezzi
• Stefano Melacci
• Alessandro Betti
• Marco Maggini
• Marco Gori

Unsupervised learning from continuous visual streams is a challenging problem that cannot be naturally and efficiently managed in the classic batch-mode setting of computation. The information stream must be carefully processed accordingly to an appropriate spatio-temporal distribution of the visual data, while most approaches of learning commonly assume uniform probability density. In this paper we focus on unsupervised learning for transferring visual information in a truly online setting by using a computational model that is inspired to the principle of least action in physics. The maximization of the mutual information is carried out by a temporal process which yields online estimation of the entropy terms. The model, which is based on second-order differential equations, maximizes the information transfer from the input to a discrete space of symbols related to the visual features of the input, whose computation is supported by hidden neurons. In order to better structure the input probability distribution, we use a human-like focus of attention model that, coherently with the information maximization model, is also based on second-order differential equations. We provide experimental results to support the theory by showing that the spatio-temporal filtering induced by the focus of attention allows the system to globally transfer more information from the input stream over the focused areas and, in some contexts, over the whole frames with respect to the unfiltered case that yields uniform probability distributions.

IJCAI Conference 2020 Conference Paper

Human-Driven FOL Explanations of Deep Learning

• Gabriele Ciravegna
• Francesco Giannini
• Marco Gori
• Marco Maggini
• Stefano Melacci

Deep neural networks are usually considered black-boxes due to their complex internal architecture, that cannot straightforwardly provide human-understandable explanations on how they behave. Indeed, Deep Learning is still viewed with skepticism in those real-world domains in which incorrect predictions may produce critical effects. This is one of the reasons why in the last few years Explainable Artificial Intelligence (XAI) techniques have gained a lot of attention in the scientific community. In this paper, we focus on the case of multi-label classification, proposing a neural network that learns the relationships among the predictors associated to each class, yielding First-Order Logic (FOL)-based descriptions. Both the explanation-related network and the classification-related network are jointly learned, thus implicitly introducing a latent dependency between the development of the explanation mechanism and the development of the classifiers. Our model can integrate human-driven preferences that guide the learning-to-explain process, and it is presented in a unified framework. Different typologies of explanations are evaluated in distinct experiments, showing that the proposed approach discovers new knowledge and can improve the classifier performance.

ECAI Conference 2020 Conference Paper

Relational Neural Machines

• Giuseppe Marra
• Michelangelo Diligenti
• Francesco Giannini
• Marco Gori
• Marco Maggini

Deep learning has been shown to achieve impressive results in several tasks where a large amount of training data is available. However, deep learning solely focuses on the accuracy of the predictions, neglecting the reasoning process leading to a decision, which is a major issue in life-critical applications. Probabilistic logic reasoning allows to exploit both statistical regularities and specific domain expertise to perform reasoning under uncertainty, but its scalability and brittle integration with the layers processing the sensory data have greatly limited its applications. For these reasons, combining deep architectures and probabilistic logic reasoning is a fundamental goal towards the development of intelligent agents operating in complex environments. This paper presents Relational Neural Machines, a novel framework allowing to jointly train the parameters of the learners and of a First–Order Logic based reasoner. A Relational Neural Machine is able to recover both classical learning from supervised data in case of pure sub-symbolic learning, and Markov Logic Networks in case of pure symbolic reasoning, while allowing to jointly train and perform inference in hybrid learning tasks. Proper algorithmic solutions are devised to make learning and inference tractable in large-scale problems. The experiments show promising results in different relational tasks.

AAAI Conference 2018 Conference Paper

Characterization of the Convex Łukasiewicz Fragment for Learning From Constraints

• Francesco Giannini
• Michelangelo Diligenti
• Marco Gori
• Marco Maggini

This paper provides a theoretical insight for the integration of logical constraints into a learning process. In particular it is proved that a fragment of the Łukasiewicz logic yields a set of convex constraints. The fragment is enough expressive to include many formulas of interest such as Horn clauses. Using the isomorphism of Łukasiewicz formulas and McNaughton functions, logical constraints are mapped to a set of linear constraints once the predicates are grounded on a given sample set. In this framework, it is shown how a collective classification scheme can be formulated as a quadratic programming problem, but the presented theory can be exploited in general to embed logical constraints into a learning process. The proposed approach is evaluated on a classification task to show how the use of the logical rules can be effective to improve the accuracy of a trained classifier.

NeSy Conference 2011 Conference Paper

Towards Developmental AI: The paradox of Ravenous Intelligent Agents

• Michelangelo Diligenti
• Marco Gori
• Marco Maggini

ECAI Conference 2010 Conference Paper

Multitask Kernel-based Learning with Logic Constraints

• Michelangelo Diligenti
• Marco Gori
• Marco Maggini
• Leonardo Rigutini

This paper presents a general framework to integrate prior knowledge in the form of logic constraints among a set of task functions into kernel machines. The logic propositions provide a partial representation of the environment, in which the learner operates, that is exploited by the learning algorithm together with the information available in the supervised examples. In particular, we consider a multi-task learning scheme, where multiple unary predicates on the feature space are to be learned by kernel machines and a higher level abstract representation consists of logic clauses on these predicates, known to hold for any input. A general approach is presented to convert the logic clauses into a continuous implementation, that processes the outputs computed by the kernel-based predicates. The learning task is formulated as a primal optimization problem of a loss function that combines a term measuring the fitting of the supervised examples, a regularization term, and a penalty term that enforces the constraints on both supervised and unsupervised examples. The proposed semi-supervised learning framework is particularly suited for learning in high dimensionality feature spaces, where the supervised training examples tend to be sparse and generalization difficult. Unlike for standard kernel machines, the cost function to optimize is not generally guaranteed to be convex. However, the experimental results show that it is still possible to find good solutions using a two stage learning schema, in which first the supervised examples are learned until convergence and then the logic constraints are forced. Some promising experimental results on artificial multi-task learning tasks are reported, showing how the classification accuracy can be effectively improved by exploiting the a priori rules and the unsupervised examples.

ECAI Conference 2006 Conference Paper

Automatic Term Categorization by Extracting Knowledge from the Web

• Leonardo Rigutini
• Ernesto Di Iorio
• Marco Ernandes
• Marco Maggini

This paper addresses the problem of categorizing terms or lexical entities into a predefined set of semantic domains exploiting the knowledge available on-line in the Web. The proposed system can be effectively used for the automatic expansion of thesauri, limiting the human effort to the preparation of a small training set of tagged entities. The classification of terms is performed by modeling the contexts in which terms from the same class usually appear. The Web is exploited as a significant repository of contexts that are extracted by querying one or more search engines. In particular, it is shown how the required knowledge can be obtained directly from the snippets returned by the search engines without the overhead of document downloads. Since the Web is continuously updated “World Wide”, this approach allows us to face the problem of open-domain term categorization handling both the geographical and temporal variability of term semantics. The performances attained by different text classifiers are compared, showing that the accuracy results are very good independently of the specific model, thus validating the idea of using term contexts extracted from search engine snippets. Moreover, the experimental results indicate that only very few training examples are needed to reach the best performance (over 90% for the F1 measure).

TCS Journal 2004 Journal Article

Neural computation, social networks, and topological spectra

• Michelangelo Diligenti
• Marco Gori
• Marco Maggini

This paper emphasizes some intriguing links between neural computation on graphical domains and social networks, like those used in nowadays search engines to score the page authority. It is pointed out that the introduction of web domains creates a unified mathematical framework for these computational schemes. It is shown that one of the major limitations of currently used connectionist models, namely their scarce ability to capture the topological features of patterns, can be effectively faced by computing the node rank according to social-based computation, like Google's PageRank. The main contribution of the paper is the introduction of a novel graph spectral notion, which can be naturally used for the graph isomorphism problem. In particular, a class of graphs is introduced for which the problem is proven to be polynomial. It is also pointed out that the derived spectral representations can be nicely combined with learning, thus opening the doors to many applications typically faced within the framework of neural computation.

IJCAI Conference 2003 Conference Paper

A Learning Algorithm for Web Page Scoring Systems

• Michelangelo Diligenti
• Marco Gori
• Marco Maggini

Hyperlink analysis is a successful approach to define algorithms which compute the relevance of a document on the basis of the citation graph. In this paper we propose a technique to learn the parameters of the page ranking model using a set of pages labeled as relevant or not relevant by a supervisor. In particular we describe a learning algorithm applied to a scheme similar to PageRank. The ranking algorithm is based on a probabilistic Web surfer model and its parameters are optimized in order to increase the probability of the surfer to visit a page labeled as relevant and to reduce it for the pages labeled as not relevant. The experimental results show the effectiveness of the proposed technique in reorganizing the page ordering in the ranking list accordingly to the examples provided in the learning set.