EDBT 2026 Demo / reviewers in the wild / expert
Francesco Giannini
dblp:198/0854
· DBLP profile ↗
39ranked-venue papers
11as first author
28since 2021 · last 2026
—ORCID · conflict
Domains — the database's venue-derived domains; a paper can count in several
Artificial intelligence and machine learning · 31 · 7 first-author · 20 since 2021Graphics, computer vision, multimedia, augmented reality and games · 7 · 1 first-author · 3 since 2021Applied, interdisciplinary, general and emerging computing · 7 · 4 first-author · 7 since 2021Software engineering, systems software and programming languages · 5 · 3 first-author · 5 since 2021Databases, data management, data science and information retrieval · 4 · 1 first-author · 1 since 2021Theory of computation · 3 · 1 first-author · 2 since 2021Human-computer interaction and ubiquitous computing · 1 · 1 since 2021
| Year | Publication | Venue | Position |
|---|---|---|---|
| 2026 | DeepProofLog: Efficient Proving in Deep Stochastic Logic ProgramsabstractNeurosymbolic (NeSy) AI combines neural architectures and symbolic reasoning to improve accuracy, interpretability, and generalization. While logic inference on top of subsymbolic modules has been shown to effectively guarantee these properties, this often comes at the cost of reduced scalability, which can severely limit the usability of NeSy models. This paper introduces DeepProofLog (DPrL), a novel NeSy system based on stochastic logic programs, which addresses the scalability limitations of previous methods. DPrL parameterizes all derivation steps with neural networks, allowing efficient neural guidance over the proving system. Additionally, we establish a formal mapping between the resolution process of our deep stochastic logic programs and Markov Decision Processes, enabling the application of dynamic programming and reinforcement learning techniques for efficient inference and learning. This theoretical connection improves scalability for complex proof spaces and large knowledge bases. Our experiments on standard NeSy benchmarks and knowledge graph reasoning tasks demonstrate that DPrL outperforms existing state-of-the-art NeSy systems, advancing scalability to larger and more complex settings than previously possible. Ying Jiao, Rodrigo Castellano Ontiveros, Luc De Raedt, Marco Gori, Francesco Giannini, Michelangelo Diligenti, Giuseppe Marra |
AAAI | 5 |
| 2026 | If Concept Bottlenecks are the Question, are Foundation Models the Answer?abstractAbstract Concept Bottleneck Models (CBMs) are neural networks designed to conjoin high performance with ante-hoc interpretability. CBMs work by first mapping inputs (e.g., images) to high-level concepts (e.g., visible objects and their properties) and then use these to solve a downstream task (e.g., tagging or scoring an image) in an interpretable manner. Their performance and interpretability, however, hinge on the quality of the concepts they learn . The go-to strategy for ensuring good quality concepts is to leverage expert annotations, which are expensive to collect and seldom available in applications. Researchers have recently addressed this issue by introducing “VLM-CBM” architectures that replace manual annotations with weak supervision from foundation models. It is however unclear what the impact of doing so is on the quality of the learned concepts. To answer this question, we put state-of-the-art VLM-CBMs to the test, analyzing their learned concepts empirically using a selection of significant metrics. Our results show that, depending on the task, VLM supervision can noticeably differ from expert annotations, and that concept accuracy and quality are not strongly correlated. Our code is available at https://github.com/debryu/CQA . Nicola Debole, Pietro Barbiero, Francesco Giannini, Andrea Passerini, Stefano Teso, Emanuele Marconato |
Mach. Learn. | 3 |
| 2025 | A distributed control architecture for logistics operations in flexible manufacturing systemsabstractIn this paper, the problem of controlling autonomous vehicles in a Flexible Manufacturing System is addressed in order to optimize logistic operations. To this end, vehicles are required to navigate between machines and from/to the Load/Unload station. The core contribution of this paper is to propose a set-theoretic distributed Model Predictive Control in charge of controlling the autonomous vehicles properly integrated with a Reinforcement Learning scheme to address the routing problem. In addition, vehicles are organized as platoons in order to improve the efficiency of the overall architecture. The numerical simulation shows the effectiveness of the proposed approach. Francesco Giannini, Domenico Famularo, Giancarlo Fortino, Giuseppe Franzè |
CoDIT | 1 |
| 2025 | Counterfactual Concept Bottleneck ModelsabstractCurrent deep learning models are not designed to simultaneously address three fundamental questions: predict class labels to solve a given classification task (the "What?"), simulate changes in the situation to evaluate how this impacts class predictions (the "How?"), and imagine how the scenario should change to result in different class predictions (the "Why not?"). While current approaches in causal representation learning and concept interpretability are designed to address some of these questions individually (such as Concept Bottleneck Models, which address both ``what'' and ``how'' questions), no current deep learning model is specifically built to answer all of them at the same time. To bridge this gap, we introduce CounterFactual Concept Bottleneck Models (CF-CBMs), a class of models designed to efficiently address the above queries all at once without the need to run post-hoc searches. Our experimental results demonstrate that CF-CBMs: achieve classification accuracy comparable to black-box models and existing CBMs (“What?”), rely on fewer important concepts leading to simpler explanations (“How?”), and produce interpretable, concept-based counterfactuals (“Why not?”). Additionally, we show that training the counterfactual generator jointly with the CBM leads to two key improvements: (i) it alters the model's decision-making process, making the model rely on fewer important concepts (leading to simpler explanations), and (ii) it significantly increases the causal effect of concept interventions on class predictions, making the model more responsive to these changes. Gabriele Dominici, Pietro Barbiero, Francesco Giannini, Martin Gjoreski, Giuseppe Marra, Marc Langheinrich |
ICLR | 3 |
| 2025 | Grounding Methods for Neural-Symbolic AIabstractA large class of Neural-Symbolic (NeSy) methods employs a machine learner to process the input entities, while relying on a reasoner based on First-Order Logic to represent and process more complex relationships among the entities. A fundamental role for these methods is played by the process of logic grounding, which determines the relevant substitutions for the logic rules using a (sub)set of entities. Some NeSy methods use an exhaustive derivation of all possible substitutions, preserving the full expressive power of the logic knowledge, but leading to a combinatorial explosion of the number of ground formulas to consider and, therefore, strongly limiting their scalability. Other methods rely on heuristic-based selective derivations, which are generally more computationally efficient, but lack a justification and provide no guarantees of preserving the information provided to and returned by the reasoner. Taking inspiration from multi-hop symbolic reasoning, this paper proposes a parametrized family of grounding methods generalizing classic Backward Chaining. Different selections within this family allow to obtain commonly employed grounding methods as special cases, and to control the trade-off between expressiveness and scalability of the reasoner. The experimental results show that the selection of the grounding criterion is often as important as the NeSy method itself. Rodrigo Castellano Ontiveros, Francesco Giannini, Marco Gori, Giuseppe Marra, Michelangelo Diligenti |
IJCAI | 2 |
| 2025 | Categorical Explaining Functors: Ensuring Coherence in Logical ExplanationsabstractPost-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. Stefano Fioravanti, Francesco Giannini, Pietro Barbiero, Paolo Frazzetto, Roberto Confalonieri 0001, Fabio Zanasi, Nicolò Navarin |
KR | 2 |
| 2025 | Distilling KGE black boxes into interpretable NeSy modelsabstractKnowledge Graph Embedding (KGE) models have shown remarkable performances in the knowledge graph completion task, thanks to their ability to capture and represent complex relational patterns. Indeed, modern KGEs encompass different inductive biases, which can account for relational patterns like reasoning compositional chains, symmetries, anti-symmetries, hierarchical patterns, etc. However, KGE models inherently lack interpretability, as their generalization capabilities are purely focused on mapping human interpretable units of information, like constants and predicates, into vector embeddings in a dense latent space, which is completely opaque to a human operator. On the other hand, different Neural-Symbolic (NeSy) methods have shown competitive results in knowledge completion tasks, but their focus on achieving high accuracy often leads to sacrificing interpretability. Many existing NeSy approaches, while inherently interpretable, resort to blending their predictions with opaque KGEs to boost performance, ultimately diminishing their explanatory power. This paper introduces a novel approach to address this limitation by applying a post-hoc NeSy method to KGE models. This strategy ensures both high fidelity to KGE models and the inherent interpretability of NeSy approaches. The proposed framework defines NeSy reasoners that generate explicit logic proofs using predefined or learned rules, ensuring transparent and explainable predictions. We evaluate the methodology using both accuracy and explainability-based metrics, demonstrating the effectiveness of our approach. Rodrigo Castellano Ontiveros, Francesco Giannini, Michelangelo Diligenti |
NeSy | 2 |
| 2025 | Deferring Concept Bottleneck Models: Learning to Defer Interventions to Inaccurate ExpertsabstractConcept Bottleneck Models (CBMs) are interpretable machine learning models that ground 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 the reasons for deferring on the final task. Our results show that DCBMs can achieve high predictive performance and interpretability by deferring only when needed. Andrea Pugnana, Riccardo Massidda, Francesco Giannini, Pietro Barbiero, Mateo Espinosa Zarlenga, Roberto Pellungrini, Gabriele Dominici, Fosca Giannotti, Davide Bacciu |
NeurIPS | 3 |
| 2025 | Relational reasoning networksabstractNeural-symbolic methods integrate neural architectures, knowledge representation and reasoning. However, they have struggled with both the intrinsic uncertainty of the observations and scaling to real-world applications. This paper presents Relational Reasoning Networks (R2N), a novel end-to-end model that performs relational reasoning in the latent space of a deep learner architecture, where the representations of constants, ground atoms and their manipulations are learned in an integrated fashion. Unlike flat architectures such as Knowledge Graph Embedders, which can only represent relations between entities, R2Ns define an additional computational structure, accounting for higher-level relations among the ground atoms. The considered relations can be explicitly known, like the ones defined by logic formulas, or defined as unconstrained correlations among groups of ground atoms. R2Ns can be applied to purely symbolic tasks or as a neural-symbolic platform to integrate learning and reasoning in heterogeneous problems with entities represented both symbolically and feature-based. The proposed model overtakes the limitations of previous neural-symbolic methods that have been either limited in terms of scalability or expressivity. The proposed methodology is shown to achieve state-of-the-art results in different experimental settings. Giuseppe Marra, Michelangelo Diligenti, Francesco Giannini |
Knowl. Based Syst. | 3 |
| 2025 | An Intelligent Multi-Layer Control Architecture for Logistics Operations of Autonomous Vehicles in Manufacturing SystemsabstractIn this paper, autonomous vehicles are considered for addressing logistic operations in manufacturing systems. The starting idea consists in organizing a given group of autonomous robots/vehicles in a finite set of platoons in charge to accomplish prescribed job(s) within the manufacturing system. Three aspects are then needed to be formally outlined: task scheduling, routing decisions and command inputs computations. Here, a new distributed multi-layer architecture has been conceived by using three methodologies: timed colored Petri nets, deep reinforcement learning and model predictive control. Roughly speaking, timed colored Petri nets are exploited to formally model the manufacturing system so that an optimal scheduling task complying with the required jobs and the available vehicles is derived; then, run-time routing decisions are obtained by using a distributed reinforcement learning algorithm which exploits the available information provided by the vehicle sensor module; finally, the distributed model predictive control algorithm is built by resorting to a set-theoretic approach where most of the computations are off-line performed. A flexible manufacturing system consisting of four machines and a Load/Unload station is used for simulation purposes. Specifically, five jobs are hypothesized and some scenarios with an increasing number of available vehicles are simulated. In order to evaluate the benefits of the proposed approach, a time criterion based on the completion of all the jobs is considered with the aim to put in light that increasing the number of vehicles improves the control performance until congestion phenomena become unavoidable. Note to Practitioners—This paper proposes an innovative methodology for addressing the logistic operations within flexible manufacturing systems (FMSs) by using a team of autonomous ground vehicles. Unlike existing approaches, the core of this framework consists in combining along a hierarchical structure the capabilities of timed colored Petri nets and the deep reinforcement learning techniques to determine a near-optimal scheduling and run-time routing decisions that are provided to the distributed model predictive units in charge to accomplish the prescribed task. This multi-layer architecture has two main merits: a single platoon, completely disconnected from the others, is devoted to perform its own job; computational burdens are affordable during the on-line operations because most of the computations are moved in the off-line phase. Domenico Famularo, Giancarlo Fortino, Francesco Pupo, Francesco Giannini, Giuseppe Franzè |
IEEE Trans Autom. Sci. Eng. | 4 |
| 2024 | A distributed control architecture for sustainable routing decisions of autonomous vehicle platoons subject to cyber attacksabstractThis paper delves into the resilience challenges encountered by platoons of self-governing agents navigating city streets, particularly focusing on the impact of misleading data infiltrating neighborly communication channels. The core issue centers around defining the trajectory paths of vehicles through routing decisions that must adhere to traffic fl ow constraints. To address this, we employ a decentralized model predictive control (MPC) strategy, leveraging deep reinforcement learning (DRL) abilities for effective decision-making in managing the regulation tasks. Additionally, the paper explores cybersecurity concerns by developing an effective anomaly detection system and practical attack countermeasures to enhance the resilience of the vehicle platoons against data tampering and other cyber threats. Through simulations and experiments, we demonstrate the efficacy of our approach in maintaining smooth traffic flow and ensuring secure communication within autonomous vehicle networks. Domenico Famularo, Francesco Giannini, Giancarlo Fortino, Giuseppe Franzè |
CoDIT | 2 |
| 2024 | Workshop on Human-Interpretable AIabstractThis workshop aims to spearhead research on Human-Interpretable Artificial Intelligence (HI-AI) by providing: (i) a general overview of the key aspects of HI-AI, in order to equip all researchers with the necessary background and set of definitions; (ii) novel and interesting ideas coming from both invited talks and top paper contributions; (iii) the chance to engage in dialogue with prominent scientists during poster presentations and coffee breaks. The workshop welcomes contributions covering novel interpretable-by-design or post-hoc approaches, as well as theoretical analysis of existing works. Additionally, we accept visionary contributions speculating on the future potential of this field. Finally, we welcome contributions from related fields such as Ethical AI, Knowledge-driven Machine learning, Human-machine Interaction, but also applications in Medicine and Industry, and analyses from Regulatory experts. Gabriele Ciravegna, Mateo Espinosa Zarlenga, Pietro Barbiero, Francesco Giannini, Zohreh Shams, Damien Garreau, Mateja Jamnik, Tania Cerquitelli |
KDD | 4 |
| 2024 | Relational Concept Bottleneck ModelsabstractThe design of interpretable deep learning models working in relational domains poses an open challenge: interpretable deep learning methods, such as Concept Bottleneck Models (CBMs), are not designed to solve relational problems, while relational deep learning models, such as Graph Neural Networks (GNNs), are not as interpretable as CBMs. To overcome these limitations, we propose Relational Concept Bottleneck Models (R-CBMs), a family of relational deep learning methods providing interpretable task predictions. As special cases, we show that R-CBMs are capable of both representing standard CBMs and message passing GNNs. To evaluate the effectiveness and versatility of these models, we designed a class of experimental problems, ranging from image classification to link prediction in knowledge graphs. In particular we show that R-CBMs (i) match generalization performance of existing relational black-boxes, (ii) support the generation of quantified concept-based explanations, (iii) effectively respond to test-time interventions, and (iv) withstand demanding settings including out-of-distribution scenarios, limited training data regimes, and scarce concept supervisions. Pietro Barbiero, Francesco Giannini, Gabriele Ciravegna, Michelangelo Diligenti, Giuseppe Marra |
NeurIPS | 2 |
| 2024 | Interpretable Concept-Based Memory ReasoningabstractThe lack of transparency in the decision-making processes of deep learning systems presents a significant challenge in modern artificial intelligence (AI), as it impairs users’ ability to rely on and verify these systems. To address this challenge, Concept Bottleneck Models (CBMs) have made significant progress by incorporating human-interpretable concepts into deep learning architectures. This approach allows predictions to be traced back to specific concept patterns that users can understand and potentially intervene on. However, existing CBMs’ task predictors are not fully interpretable, preventing a thorough analysis and any form of formal verification of their decision-making process prior to deployment, thereby raising significant reliability concerns. To bridge this gap, we introduce Concept-based Memory Reasoner (CMR), a novel CBM designed to provide a human-understandable and provably-verifiable task prediction process. Our approach is to model each task prediction as a neural selection mechanism over a memory of learnable logic rules, followed by a symbolic evaluation of the selected rule. The presence of an explicit memory and the symbolic evaluation allow domain experts to inspect and formally verify the validity of certain global properties of interest for the task prediction process. Experimental results demonstrate that CMR achieves better accuracy-interpretability trade-offs to state-of-the-art CBMs, discovers logic rules consistent with ground truths, allows for rule interventions, and allows pre-deployment verification. David Debot, Pietro Barbiero, Francesco Giannini, Gabriele Ciravegna, Michelangelo Diligenti, Giuseppe Marra |
NeurIPS | 3 |
| 2024 | Embedding the State Trajectories of Nonlinear Systems via Multimodel Linear Descriptions: A Data-Driven-Based AlgorithmabstractIn this article, the problem of generating multimodel state space descriptions in a data-driven context to embed the dynamic behavior of nonlinear systems is addressed. The proposed methodology takes advantage of three ingredients: 1) linear time-invariant system behavior; 2) data-driven modeling; and 3) reinforcement learning (RL) technicalities. These elements are properly combined to develop a data-driven algorithm capable to derive an accurate outer convex approximation of the nonlinear evolution. In particular, an actor-critic RL scheme is designed to efficiently comply with the exhaustive research on the whole parameter space. At each iteration, the effectiveness of the obtained uncertain polytopic model is tested by a probabilistic approach based on a confidence level metrics. As the main merits of the proposed approach are concerned, the following aspect clearly stands up: the development of an interdisciplinary methodology that takes advantage of system theory, probabilistic arguments and RL capabilities giving rise to an harmonized architecture in charge to deal with a vast class of nonlinear systems. Finally, the validity of the proposed approach is tested by resorting to benchmark examples that allow to quantify the level of accuracy of the computed convex hull. Giuseppe Franzè, Francesco Giannini, Vicenç Puig, Giancarlo Fortino |
IEEE Trans. Syst. Man Cybern. Syst. | 2 |
| 2023 | A Neural Network and Model Predictive Control Based Resilient Architecture for Constrained Cyber-Physical SystemsabstractIn this paper, the resilient control problem for constrained cyber-physical systems subject to stealthy data intrusions on the communication channels is considered. The key idea consists in designing a neural network to act as the anomaly detector during the on-line operations. Accordingly the controller unit, based on model predictive control arguments, is developed to take advantage of the resulting detection capabilities. As its main merits are concerned, the overall control architecture has a two-fold merit with respect to the existing literature: it is avoided the need of modifying the detector structure whenever a different class of attacks is considered, and the occurrence of false positive events is significantly mitigated. Finally, a numerical example is provided to show the effectiveness and peculiarities of the proposed approach. Luigi D'Alfonso, Giuseppe Franzè, Francesco Giannini, Francesco Tedesco |
CoDIT | 3 |
| 2023 | A Set-Theoretic Receding Horizon Control Based on a Q-Learning Approach for Sustainability PurposesabstractThis paper presents a set-theoretic receding horizon control strategy for platoons of autonomous vehicles driving in smart cities context. In order to reduce traffic and$CO_{2}$emissions, we propose a path planer based on Deep Reinforcement Learning (DRL). The advantages of this solution is the ability to deal with the actual traffic congestion, while driving the autonomous vehicles to their destination and fulfilling the constraints. In particular, the high-level routing decisions are translated into set-points for the receding horizon controllers, making the control actions on the vehicle dynamics more computational efficient. In order to show the effectiveness of the overall architecture, a campaign of simulations on a platoon of eight vehicles, moving in the city center of Bologna in Italy, is provided. Francesco Giannini, Giuseppe Franzè, Francesco Pupo, Giancarlo Fortino |
CoDIT | 1 |
| 2023 | Interpretable Neural-Symbolic Concept ReasoningabstractDeep learning methods are highly accurate, yet their opaque decision process prevents them from earning full human trust. Concept-based models aim to address this issue by learning tasks based on a set of human-understandable concepts. However, state-of-the-art concept-based models rely on high-dimensional concept embedding representations which lack a clear semantic meaning, thus questioning the interpretability of their decision process. To overcome this limitation, we propose the Deep Concept Reasoner (DCR), the first interpretable concept-based model that builds upon concept embeddings. In DCR, neural networks do not make task predictions directly, but they build syntactic rule structures using concept embeddings. DCR then executes these rules on meaningful concept truth degrees to provide a final interpretable and semantically-consistent prediction in a differentiable manner. Our experiments show that DCR: (i) improves up to +25% w.r.t. state-of-the-art interpretable concept-based models on challenging benchmarks (ii) discovers meaningful logic rules matching known ground truths even in the absence of concept supervision during training, and (iii), facilitates the generation of counterfactual examples providing the learnt rules as guidance. Pietro Barbiero, Gabriele Ciravegna, Francesco Giannini, Mateo Espinosa Zarlenga, Lucie Charlotte Magister, Alberto Paolo Tonda, Pietro Liò, Frédéric Precioso, Mateja Jamnik, Giuseppe Marra |
ICML | 3 |
| 2023 | Enhancing Embedding Representations of Biomedical Data using Logic KnowledgeabstractKnowledge Graph Embeddings (KGE) have become a quite popular class of models specifically devised to deal with ontologies and graph structure data, as they can implicitly encode statistical dependencies between entities and relations in a latent space. KGE techniques are particularly effective for the biomedical domain, where it is quite common to deal with large knowledge graphs underlying complex interactions between biological and chemical objects. Recently in the literature, the PharmKG dataset has been proposed as one of the most challenging knowledge graph biomedical benchmark, with hundreds of thousands of relational facts between genes, diseases and chemicals. Despite KGEs can scale to very large relational domains, they generally fail at representing more complex relational dependencies between facts, like logic rules, which may be fundamental in complex experimental settings. In this paper, we exploit logic rules to enhance the embedding representations of KGEs on the PharmKG dataset. To this end, we adopt Relational Reasoning Network (R2N), a recently proposed neural-symbolic approach showing promising results on knowledge graph completion tasks. An R2N uses the available logic rules to build a neural architecture that reasons over KGE latent representations. In the experiments, we show that our approach is able to significantly improve the current state-of-the-art on the PharmKG dataset. Finally, we provide an ablation study to experimentally compare the effect of alternative sets of rules according to different selection criteria and varying the number of considered rules. Michelangelo Diligenti, Francesco Giannini, Stefano Fioravanti, Caterina Graziani, Moreno Falaschi, Giuseppe Marra |
IJCNN | 2 |
| 2023 | Linguistic Feature Injection for Efficient Natural Language ProcessingabstractTransformers have been established as one of the most effective neural approach in performing various Natural Language Processing tasks. However, following common trend in modern deep architectures, their scale has quickly grown to an extent that reduces the concrete possibility for several enterprises to train such models from scratch. Indeed, despite their high-level performances, Transformers have the general drawback of requiring a huge amount of training data, computational resources and energy consumption to be successfully optimized. For this reason, more recent architectures like Bidirectional Encoder Representations from Transformers rely on unlabeled data to pre-train the model, which is later fine-tuned for a specific downstream task using a relatively smaller amount of training data. In a similar fashion, this paper considers a plug-and-play framework that can be used to inject multiple syntactic features, like Part-of-Speech Tagging or Dependency Parsing, into any kind of pre-trained Transformer. This novel approach allows to perform sequence-to-sequence labeling tasks by exploiting: (i) the (more abundant) available training data that is also used to learn the syntactic features, (ii) the language data that is used to pre-train the transformer model. The experimental results show that our approach improves over the baseline performances of the underlying model in different datasets, thus proving the effectiveness of employing syntactic language information for semantic regularization. In addition, we show that our architecture has a huge efficiency advantage over pure large language models. Indeed, by using a model with limited size, but whose input data are enriched with syntactic information, we show that it is possible to obtain a significant reduction of CO2 emissions without decreasing the prediction performances. Stefano Fioravanti, Andrea Zugarini, Francesco Giannini, Leonardo Rigutini, Marco Maggini, Michelangelo Diligenti |
IJCNN | 3 |
| 2023 | Interpretable Graph Networks Formulate Universal Algebra ConjecturesabstractThe rise of Artificial Intelligence (AI) recently empowered researchers to investigate hard mathematical problems which eluded traditional approaches for decades. Yet, the use of AI in Universal Algebra (UA)---one of the fields laying the foundations of modern mathematics---is still completely unexplored.
This work proposes the first use of AI to investigate UA's conjectures with an equivalent equational and topological characterization. While topological representations would enable the analysis of such properties using graph neural networks, the limited transparency and brittle explainability of these models hinder their straightforward use to empirically validate existing conjectures or to formulate new ones.
To bridge these gaps, we propose a general algorithm generating AI-ready datasets based on UA's conjectures, and introduce a novel neural layer to build fully interpretable graph networks. The results of our experiments demonstrate that interpretable graph networks: (i) enhance interpretability without sacrificing task accuracy, (ii) strongly generalize when predicting universal algebra's properties, (iii) generate simple explanations that empirically validate existing conjectures, and (iv) identify subgraphs suggesting the formulation of novel conjectures. Francesco Giannini, Stefano Fioravanti, Oguzhan Keskin, Alisia Maria Lupidi, Lucie Charlotte Magister, Pietro Liò, Pietro Barbiero |
NeurIPS | 1 |
| 2023 | Logic Explained Networks
Gabriele Ciravegna, Pietro Barbiero, Francesco Giannini, Marco Gori, Pietro Liò, Marco Maggini, Stefano Melacci |
Artif. Intell. | 3 |
| 2023 | T-norms driven loss functions for machine learningabstractAbstract Injecting prior knowledge into the learning process of a neural architecture is one of the main challenges currently faced by the artificial intelligence community, which also motivated the emergence of neural-symbolic models. One of the main advantages of these approaches is their capacity to learn competitive solutions with a significant reduction of the amount of supervised data. In this regard, a commonly adopted solution consists of representing the prior knowledge via first-order logic formulas, then relaxing the formulas into a set of differentiable constraints by using a t-norm fuzzy logic. This paper shows that this relaxation, together with the choice of the penalty terms enforcing the constraint satisfaction, can be unambiguously determined by the selection of a t-norm generator, providing numerical simplification properties and a tighter integration between the logic knowledge and the learning objective. When restricted to supervised learning, the presented theoretical framework provides a straight derivation of the popular cross-entropy loss, which has been shown to provide faster convergence and to reduce the vanishing gradient problem in very deep structures. However, the proposed learning formulation extends the advantages of the cross-entropy loss to the general knowledge that can be represented by neural-symbolic methods. In addition, the presented methodology allows the development of novel classes of loss functions, which are shown in the experimental results to lead to faster convergence rates than the approaches previously proposed in the literature. Francesco Giannini, Michelangelo Diligenti, Marco Maggini, Marco Gori, Giuseppe Marra |
Appl. Intell. | 1 |
| 2023 | A Sustainable Multi-Agent Routing Algorithm for Vehicle Platoons in Urban NetworksabstractIn this paper, a sustainable routing algorithm for vehicle platoons operating in smart urban networks is presented. The proposed approach makes use of deep reinforcement learning (DRL) and set-theoretic model predictive control (MPC). In particular, the learning process aims at reducing traffic congestion and$CO_{2}$emissions, whereas the MPC unit allows to adequately track the assigned path by using real-time traffic data. To adequately analyze the performance of the resulting control architecture, the SUMO and MATLAB environments are used to implement complex operating scenarios where road maps data and vehicle state trajectories can be shared and exchanged. Finally, numerical studies are provided by resorting to the SUMO environment and considering a platoon of five vehicles. The resulting simulation campaign puts in light the capability of the training process to significantly mitigate the$CO_{2}$emissions of the whole platoon: from a minimum of 3.7 % to a maximum of 13% with respect to the use of the well-known Dijkstra algorithm. Francesco Giannini, Giuseppe Franzè, Francesco Pupo, Giancarlo Fortino |
IEEE Trans. Intell. Transp. Syst. | 1 |
| 2022 | Entropy-Based Logic Explanations of Neural NetworksabstractExplainable artificial intelligence has rapidly emerged since lawmakers have started requiring interpretable models for safety-critical domains. Concept-based neural networks have arisen as explainable-by-design methods as they leverage human-understandable symbols (i.e. concepts) to predict class memberships. However, most of these approaches focus on the identification of the most relevant concepts but do not provide concise, formal explanations of how such concepts are leveraged by the classifier to make predictions. In this paper, we propose a novel end-to-end differentiable approach enabling the extraction of logic explanations from neural networks using the formalism of First-Order Logic. The method relies on an entropy-based criterion which automatically identifies the most relevant concepts. We consider four different case studies to demonstrate that: (i) this entropy-based criterion enables the distillation of concise logic explanations in safety-critical domains from clinical data to computer vision; (ii) the proposed approach outperforms state-of-the-art white-box models in terms of classification accuracy. Pietro Barbiero, Gabriele Ciravegna, Francesco Giannini, Pietro Liò, Marco Gori, Stefano Melacci |
AAAI | 3 |
| 2022 | Path planning for vehicle platoons under routing decisions: a distributed approach combining Deep Reinforcement Learning and Model Predictive ControlabstractIn this paper, the path planning problem under routing decisions is considered for platoons of autonomous vehicles moving in urban road networks. The key idea is to exploit and adequately combine arguments coming from two research fields: deep reinforcement learning and model predictive control. Along these lines, a novel control architecture is proposed and its feasibility formally proved. In particular, the high-level routing decisions arising from the distributed deep reinforcement learning operations are translated into manipulable set-points for the underlying bank of receding horizon controllers by making more computational affordable and efficient the action on the vehicle dynamics. Finally, some simulations on platoon, consisting of three agents described by double-integrator models, are provided to show the effectiveness of the overall architecture. Francesco Giannini, Giancarlo Fortino, Giuseppe Franzè, Francesco Pupo |
CoDIT | 1 |
| 2022 | Extending Logic Explained Networks to Text ClassificationabstractRecently, Logic Explained Networks (LENs) have been proposed as explainable-by-design neural models providing logic explanations for their predictions.However, these models have only been applied to vision and tabular data, and they mostly favour the generation of global explanations, while local ones tend to be noisy and verbose.For these reasons, we propose LEN p , improving local explanations by perturbing input words, and we test it on text classification.Our results show that (i) LEN p provides better local explanations than LIME in terms of sensitivity and faithfulness, and (ii) logic explanations are more useful and user-friendly than feature scoring provided by LIME as attested by a human survey. Gabriele Ciravegna, Pietro Barbiero, Francesco Giannini, Davide Buffelli, Pietro Liò |
EMNLP | 4 |
| 2022 | Concept Embedding Models: Beyond the Accuracy-Explainability Trade-OffabstractDeploying AI-powered systems requires trustworthy models supporting effective human interactions, going beyond raw prediction accuracy. Concept bottleneck models promote trustworthiness by conditioning classification tasks on an intermediate level of human-like concepts. This enables human interventions which can correct mispredicted concepts to improve the model's performance. However, existing concept bottleneck models are unable to find optimal compromises between high task accuracy, robust concept-based explanations, and effective interventions on concepts---particularly in real-world conditions where complete and accurate concept supervisions are scarce. To address this, we propose Concept Embedding Models, a novel family of concept bottleneck models which goes beyond the current accuracy-vs-interpretability trade-off by learning interpretable high-dimensional concept representations. Our experiments demonstrate that Concept Embedding Models (1) attain better or competitive task accuracy w.r.t. standard neural models without concepts, (2) provide concept representations capturing meaningful semantics including and beyond their ground truth labels, (3) support test-time concept interventions whose effect in test accuracy surpasses that in standard concept bottleneck models, and (4) scale to real-world conditions where complete concept supervisions are scarce. Mateo Espinosa Zarlenga, Pietro Barbiero, Gabriele Ciravegna, Giuseppe Marra, Francesco Giannini, Michelangelo Diligenti, Zohreh Shams, Frédéric Precioso, Stefano Melacci, Adrian Weller, Pietro Liò, Mateja Jamnik |
NeurIPS | 5 |
| 2020 | A Constraint-Based Approach to Learning and ExplanationabstractIn 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 sub-portions 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. Gabriele Ciravegna, Francesco Giannini, Stefano Melacci, Marco Maggini, Marco Gori |
AAAI | 2 |
| 2020 | Relational Neural MachinesabstractDeep 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. Giuseppe Marra, Michelangelo Diligenti, Francesco Giannini, Marco Gori, Marco Maggini |
ECAI | 3 |
| 2020 | Human-Driven FOL Explanations of Deep LearningabstractDeep 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. Gabriele Ciravegna, Francesco Giannini, Marco Gori, Marco Maggini, Stefano Melacci |
IJCAI | 2 |
| 2019 | Conditions for Unnecessary Logical Constraints in Kernel Machines
Francesco Giannini, Marco Maggini |
ICANN (2) | 1 |
| 2019 | Constraint-Based Visual Generation
Giuseppe Marra, Francesco Giannini, Michelangelo Diligenti, Marco Gori |
ICANN (3) | 2 |
| 2019 | On the Relation Between Loss Functions and T-Norms
Francesco Giannini, Giuseppe Marra, Michelangelo Diligenti, Marco Maggini, Marco Gori |
ILP | 1 |
| 2019 | LYRICS: A General Interface Layer to Integrate Logic Inference and Deep Learning
Giuseppe Marra, Francesco Giannini, Michelangelo Diligenti, Marco Gori |
ECML/PKDD (2) | 2 |
| 2019 | Integrating Learning and Reasoning with Deep Logic Models
Giuseppe Marra, Francesco Giannini, Michelangelo Diligenti, Marco Gori |
ECML/PKDD (2) | 2 |
| 2019 | On a Convex Logic Fragment for Learning and ReasoningabstractIn this paper, we introduce the convex fragment of Łukasiewicz logic and discuss its possible applications in different learning schemes. The provided theoretical results are highly general because they can be exploited in any learning framework involving logical constraints. The method is of particular interest since the fragment guarantees to deal with convex constraints, which are shown to be equivalent to a set of linear constraints. Within this framework, we are able to formulate learning with kernel machines as well as collective classification as a quadratic programming problem. Francesco Giannini, Michelangelo Diligenti, Marco Gori, Marco Maggini |
IEEE Trans. Fuzzy Syst. | 1 |
| 2018 | Characterization of the Convex Łukasiewicz Fragment for Learning From Constraints
Francesco Giannini, Michelangelo Diligenti, Marco Gori, Marco Maggini |
AAAI | 1 |
| 2017 | Learning Łukasiewicz Logic Fragments by Quadratic Programming
Francesco Giannini, Michelangelo Diligenti, Marco Gori, Marco Maggini |
ECML/PKDD (1) | 1 |