Context: The adoption of Artificial Intelligence (AI) in industrial production systems has raised significant expectations for increased efficiency and innovation. Nevertheless, challenges such as the distributed nature of industrial operations, the heterogeneity of physical devices, and the complexity of real-world processes continue to hinder AI integration. Digital Twins (DTs) have emerged as a promising abstraction to decouple physical complexity from digital representations, facilitating more effective system management. Objective: This work investigates how AI can be systematically integrated with DTs in industrial contexts. The goal is to identify and characterize a set of interaction patterns that leverage the complementary strengths of AI and DTs to enhance industrial intelligence and performance. Methods: Drawing on a structured view of how responsibilities can be shared between AI technologies and DT-enabled shop floors, the paper defines four interaction patterns—AI Observing DTs, AI Advising DTs, AI Controlling DTs, and AI Embedded in DT. Each pattern is analyzed in terms of its roles, data and control flows, and typical application scenarios, and is illustrated on a DT-enabled physical micro-factory that reproduces realistic production conditions. Results: The four patterns show how different placements and responsibilities of AI components with respect to DT layers impact modularity, reuse of AI models, maintainability, and integration with legacy industrial systems. The micro-factory illustration highlights how the patterns can support practical use cases, including root-cause analysis of performance degradation, machine-level health monitoring, and AI-based production scheduling. Conclusion: Structuring AI–DT integration around interaction patterns provides a concrete way to bridge the gap between conceptual opportunities and operational industrial systems. The proposed patterns offer a reusable design vocabulary for positioning AI with respect to DT layers in cyber–physical production systems, and for reasoning about the architectural trade-offs of alternative integration strategies.

Planetary remote sensing missions are critical for advancing our understanding of extraterrestrial systems. They operate in highly uncertain environments where reliability and resolution are not always guaranteed, often compromising data analysis and scientific outcomes. In this paper, we consider the challenging task of estimating the quality of the signal acquired by MARSIS, the subsurface sounder aboard ESA’s Mars Express mission, which aims to map the presence of liquid water beneath the Martian surface. Quality estimation has a strategic impact on the scheduling of MARSIS observations, since the radar operates with strict constraints that greatly limit the number and size of observation opportunities available per day. Thus, maximizing the quality of scheduled observations becomes a crucial factor in reducing resource utilization and increasing the coverage of the target areas in search of liquid water. To this end, in a previous research we proposed a predict-then-optimize approach, which included a neural network regressor to predict signal quality achievable by future observation opportunities, based on contextual features. In this work, we advance the methodology by applying explainable artificial intelligence techniques that allow domain experts to interpret the results, by enhancing the comprehension of the physical phenomena that have an impact on signal acquisition. Specifically, we applied a SHAP analysis to the neural network predictions and trained an Explainable Boosting Machine (EBM) to provide interpretable models. We then analyzed and compared the results with existing domain knowledge, uncovering promising new avenues for investigation and highlighting limitations in the current dataset construction.

Unfair clause detection is an extremely useful AI application for consumer protection. Artificial intelligence has recently been successful in building systems capable to automatically detect unfair clauses in Terms of Service, and also to identify their unfairness categories. Since Large Language Models (LLMs) are nowadays bringing a revolution to the field of artificial intelligence, and in particular to natural language processing and understanding, in this paper we compare several different prompt strategies for LLMs with more traditional BERT-based fine-tuned models. Our extensive experimental evaluation aims to investigate whether it is worth using LLMs also for this challenging domain-specific task.

One of the goals of neuro-symbolic artificial intelligence is to exploit background knowledge to improve the performance of learning tasks. However, most of the existing frameworks focus on the simplified scenario where knowledge does not change over time and does not cover the temporal dimension. In this work we consider the much more challenging problem of knowledge-driven sequence classification where different portions of knowledge must be employed at different timesteps, and temporal relations are available. Our extensive experimental evaluation compares multi-stage neuro-symbolic and neural-only architectures, and it is conducted on a newly-introduced benchmarking framework. Results not only demonstrate the challenging nature of this novel setting, but also highlight under-explored shortcomings of neuro-symbolic methods, representing a precious reference for future research.

Modern industrial systems, characterised by distributed and fragmented equipment, present challenges due to their inherent heterogeneity and complexity. This should not impact the stakeholders' business logic, who are more concerned with the information itself rather than how it is collected or processed. Recently, Digital Twins - software copies of physical assets and systems - emerged as a pivotal strategy to bridge the cyber-physical world into an effective digital layer decoupling applications from the management and interaction with physical assets. Fostering this vision, we propose a structured industrial Digital Twins ecosystem exploiting twin relationships and hier-archies to build a digitalised replica of the whole manufacturing system structure enabling a simplified navigation and interaction with the physical world and the data it generates. To support the depicted visions, a fully functioning prototype has been implemented and evaluated in an experimental scenario.

Local Energy Communities are becoming key actors in the panorama of sustainable development. One of the biggest challenges for such communities is to become self-efficient, thanks to an efficient management of the balancing between produced and consumed energy. In order to achieve this goal, it is necessary to design and implement forecasting models that can provide accurate estimates to be subsequently used by optimization and planning algorithms. In this work, we show how neural networks, and in particular long short-term memory networks, can be used to this aim, highlighting an interesting trade-off between the computational requirements and the forecasting accuracy induced by learning different models for clusters of users.

We explore the use of residual networks and neural attention for multiple argument mining tasks. We propose a residual architecture that exploits attention, multi-task learning, and makes use of ensemble, without any assumption on document or argument structure. We present an extensive experimental evaluation on five different corpora of user-generated comments, scientific publications, and persuasive essays. Our results show that our approach is a strong competitor against state-of-the-art architectures with a higher computational footprint or corpus-specific design, representing an interesting compromise between generality, performance accuracy and reduced model size.

Recent advances in robotics have allowed the introduction of robots assisting and working together with human subjects. To promote their use and diffusion, intuitive and user-friendly interaction means should be adopted. In particular, gestures have become an established way to interact with robots since they allow to command them in an intuitive manner. In this article, we focus on the problem of gesture recognition in human–robot interaction (HRI). While this problem has been largely studied in the literature, it poses specific constraints when applied to HRI. We propose a framework consisting in a pipeline devised to take into account these specific constraints. We implement the proposed pipeline considering, as an example, an evaluation use case. To this end, we consider standard machine learning algorithms for the classification stage and evaluate their performance considering different performance metrics for a thorough assessment.

This work investigates the problem of 6D pose estimation and robot bin picking of non-Lambertian reflecting objects based on a low-cost commercial 3D sensor. In particular, we address the task of estimating the pose of small metal hydraulic components of the same type, randomly placed in a bin. The system consists of a robot arm and an RGB-D sensor in eye-in-hand configuration. The proposed method works in two main phases. In the first phase a Convolutional Neural Network (CNN) extracts the bounding boxes of the objects contained in the bin from a single RGB image of the environment. In the second phase the 6D pose of the objects is estimated using a dense 3D reconstruction of the scene and by applying a template matching algorithm from multiple virtual views of the object CAD model. Experimental results have been carried out on a dataset containing both RGB and depth images. Preliminary experiments are also reported in the real setup.

The increasing complexity and unpredictability of many ICT scenarios will represent a major challenge for future intelligent systems.The capability to dynamically and autonomously adapt to evolving and novel situations, with a partial or limited knowledge of the domain, both at the level of individual components and at the collective level, will become a crucial need for smart devices acting in many application domains.In this paper, we envision future systems able to selfdevelop mental models of themselves and of the environment they act in.Key properties will include: learning models of own capabilities; learning how to act purposefully towards the achievement of specific goals; and learning how to act in the presence of others, i.e., at the collective level.In our work, we will introduce the vision of self-development in ICT systems, by framing its key concepts and by illustrating suitable application domains.Then, we overview the many research areas that are contributing or can potentially contribute to the realisation of the vision, and identify some key research challenges.

AMICA is an argument mining-based search engine, specifically designed for the analysis of scientific literature related to Covid-19. AMICA retrieves scientific papers based on matching keywords and ranks the results based on the papers' argumentative content. An experimental evaluation conducted on a case study in collaboration with the Italian National Institute of Health shows that the AMICA ranking agrees with expert opinion, as well as, importantly, with the impartial quality criteria indicated by Cochrane Systematic Reviews.

We present a study aimed at testing the CLAUDETTE system’s ability to generalise the concept of unfairness in consumer contracts across diverse market sectors. The data set includes 142 terms of services grouped in five sub-sets: travel and accommodation, games and entertainment, finance and payments, health and well-being, and the more general others. Preliminary results show that the classifier has satisfying performance on all the sectors.

The estimation of OD flows from mobile phone and GPS positioning data is an important application that can naturally support urban and transport studies. In this work, we first present an approach to generate OD matrices from mobile phone positioning and GPS data, and scale them with traffic counts. Then we compare these matrices, that we consider as an estimate of the potential demand for public transport, with matrices describing the actual routes of public transportation services, that represent the supply. Finally, we present a data driven approach to identifywhereandwhenthe demand for transport is not satisfied. We run experiments with different mobility datasets and the actual public transportation routes in a mid-sized Italian city. In this scenario, our approach allows to detect similar areas of unmatched demand with both such datasets. In particular, two case studies show that the proposed methodology is able to identify two existing bus lines that were recently introduced by the public transport company and local government. Finally, we show an upper bound for the reduced impact of$CO_{2}$emissions, if the unmet demand for transport is entirely shifted to public transport.

Attention is an increasingly popular mechanism used in a wide range of neural architectures. The mechanism itself has been realized in a variety of formats. However, because of the fast-paced advances in this domain, a systematic overview of attention is still missing. In this article, we define a unified model for attention architectures in natural language processing, with a focus on those designed to work with vector representations of the textual data. We propose a taxonomy of attention models according to four dimensions: the representation of the input, the compatibility function, the distribution function, and the multiplicity of the input and/or output. We present the examples of how prior information can be exploited in attention models and discuss ongoing research efforts and open challenges in the area, providing the first extensive categorization of the vast body of literature in this exciting domain.

We study annotation projection in text classification problems where source documents are published in multiple languages and may not be an exact translation of one another.In particular, we focus on the detection of unfair clauses in privacy policies and terms of service.We present the first English-German parallel asymmetric corpus for the task at hand.We study and compare several language-agnostic sentence-level projection methods.Our results indicate that a combination of word embeddings and dynamic time warping performs best.

Recent years have been tainted by market practices that continuously expose us, as consumers, to new risks and threats. We have become accustomed, and sometimes even resigned, to businesses monitoring our activities, examining our data, and even meddling with our choices. Artificial Intelligence (AI) is often depicted as a weapon in the hands of businesses and blamed for allowing this to happen. In this paper, we envision a paradigm shift, where AI technologies are brought to the side of consumers and their organizations, with the aim of building an efficient and effective counter-power. AI-powered tools can support a massive-scale automated analysis of textual and audiovisual data, as well as code, for the benefit of consumers and their organizations. This in turn can lead to a better oversight of business activities, help consumers exercise their rights, and enable the civil society to mitigate information overload. We discuss the societal, political, and technological challenges that stand before that vision.

Road traffic safety is one of the major challenges for the future of smart cities and transportation networks. Despite several solutions exist to reduce the number of fatalities and severe accidents happening daily in our roads, this reduction is smaller than expected and new methods and intelligent systems are needed. The emergency Call is an initiative of the European Commission aimed at providing rapid assistance to motorists thanks to the implementation of a unique emergency number. In this work, we study the problem of classifying the severity of accidents involving Powered Two Wheelers, by exploiting machine learning systems based on features that could be reasonably collected at the moment of the accident. An extended study on the set of features allows to identify the most important factors that enable to distinguish accident severity. The system we develop achieves around 90% of precision and recall on a large, publicly available corpus, using only a set of eleven features.

Consumer contracts often contain unfair clauses, in apparent violation of the relevant legislation.In this paper we present a new methodology for evaluating such clauses in online Terms of Services.We expand a set of tagged documents (terms of service), with a structured corpus where unfair clauses are liked to a knowledge base of rationales for unfairness, and experiment with machine learning methods on this expanded training set.Our experimental study is based on deep neural networks that aim to combine learning and reasoning tasks, one major example being Memory Networks.Preliminary results show that this approach may not only provide reasons and explanations to the user, but also enhance the automated detection of unfair clauses.

Long short-term memory (LSTM) networks have recently shown remarkable performance in several tasks that are dealing with natural language generation, such as image captioning or poetry composition. Yet, only few works have analyzed text generated by LSTMs in order to quantitatively evaluate to which extent such artificial texts resemble those generated by humans. We compared the statistical structure of LSTM-generated language to that of written natural language, and to those produced by Markov models of various orders. In particular, we characterized the statistical structure of language by assessing word-frequency statistics, long-range correlations, and entropy measures. Our main finding is that while both LSTM- and Markov-generated texts can exhibit features similar to real ones in their word-frequency statistics and entropy measures, LSTM-texts are shown to reproduce long-range correlations at scales comparable to those found in natural language. Moreover, for LSTM networks, a temperature-like parameter controlling the generation process shows an optimal value-for which the produced texts are closest to real language-consistent across different statistical features investigated.

Argument-based decision making has been employed to support a variety of reasoning tasks over medical knowledge. These include evidence-based justifications of the effects of treatments, the detection of conflicts in the knowledge base, and the enabling of uncertain and defeasible reasoning in the health-care sector. However, a common limitation of these approaches is that they rely on structured input information. Recent advances in argument mining have shown increasingly accurate results in detecting argument components and predicting their relations from unstructured, natural language texts. In this study, we discuss evidence and claim detection from Randomized Clinical Trials. To this end, we create a new annotated dataset about four different diseases (glaucoma, diabetes, hepatitis B, and hypertension), containing 976 argument components (697 containing evidence, 279 claims). Empirical results are promising, and show the portability of the proposed approach over different branches of medicine.

Artificial Intelligence and Law is undergoing a critical transformation. Traditionally focused on the development of expert systems and on a scholarly effort to develop theories and methods for knowledge representation and reasoning in the legal domain, this discipline is now adapting to a sudden change of scenery. No longer confined to the walls of academia, it has welcomed new actors, such as businesses and companies, who are willing to play a major role and seize new opportunities offered by the same transformational impact that recent AI breakthroughs are having on many other areas. As it happens, commercial interests create new opportunities but they also represent a potential threat to consumers, as the balance of power seems increasingly determined by the availability of data. We believe that while this transformation is still in progress, time is ripe for the next frontier of this field of study, where a new shift of balance may be enabled by tools and services that can be of service not only to businesses but also to consumers and, more generally, the civil society. We call that frontier consumer-empowering AI.

Two years after its entry into force, the EU General Data Protection Regulation became applicable on the 25th May 2018. Despite the long time for preparation, privacy policies of online platforms and services still often fail to comply with information duties and the standard of lawfulness of data processing. In this paper we present a new methodology for processing privacy policies under GDPR's provisions, and a novel annotated corpus, to be used by machine learning systems to automatically check the compliance and adequacy of privacy policies. Preliminary results confirm the potential of the methodology.

The crucial role played by social interactions between smart objects in the Internet of Things (IoT) is being rapidly recognized by the social IoT (SIoT) vision. In this paper, we build upon the recently introduced vision of Speaking Objects-“things” interacting through argumentation-to show how different forms of human dialogue naturally fit cooperation and coordination requirements of the SIoT. In particular, we show how Speaking Objects can exchange arguments in order to seek for information, negotiate over an issue, persuade others, deliberate actions, and so on, namely, striving to reach consensus about the state of affairs and their goals. In this context, we illustrate how argumentation naturally enables such a form of conversational coordination through practical examples and a case study scenario.

Deep networks have been successfully applied to a wide range of tasks in artificial intelligence, and game playing is certainly not an exception. In this paper, we present an experimental study to assess whether purely subsymbolic systems, such as deep networks, are capable of learning to play by the rules, without anya prioriknowledge neither of the game, nor of its rules, but only by observing the matches played by another player. Similar problems arise in many other application domains, where the goal is to learn rules, policies, behaviors, or decisions, simply by the observation of the dynamics of a system. We present a case study conducted with residual networks on the popular board game ofNine Men's Morris, showing that this kind of subsymbolic architecture is capable of correctly discriminating legal from illegal decisions, just from the observation of past matches of a single player.

Traffic simulators are effective tools to support decisions in urban planning systems, to identify criticalities, to observe emerging behaviours in road networks and to configure road infrastructures, such as road side units and traffic lights. Clearly the more realistic the simulator the more precise the insight provided to decision makers. This paper provides a first step toward the design and calibration of traffic micro-simulator to produce realistic behaviour. The long term idea is to collect and analyse real traffic traces collecting vehicular information, to cluster them in groups representing similar driving behaviours and then to extract from these clusters relevant parameters to tune the microsimulator. In this paper we have run controlled experiments where traffic traces have been synthetized to obtain different driving styles, so that the effectiveness of the clustering algorithm could be checked on known labels. We describe the overall methodology and the results already achieved on the controlled experiment, showing the clusters obtained and reporting guidelines for future experiments.

In this paper we sketch a vision of future environments densely populated by smart sensors and actuators - possibly embedded in everyday objects - that, rather than simply producing streams of data, are capable of understanding and reporting, via factual assertions and arguments, about what is happening (for sensors) and about what they can make possibly happen (for actuators). These "speaking objects" form the nodes of a dense distributed computing infrastructure that can be exploited to monitor and control activities in our everyday environment. However, the nature of speaking objects will dramatically change the approaches to implementing and coordinating the activities of distributed processes. In fact, distributed coordination is likely to become associated with the capability of argumenting about situations and about the current "state of the affairs", with the aim of triggering and directing proper distributed "conversations" to collectively reach a future desirable state. Accordingly, we discuss how such a novel vision can build upon some readily available technologies, and the research challenges that it poses. Two case studies are used as exemplary scenarios.

Consumer contracts too often present clauses that are potentially unfair to the subscriber. We present an experimental study where machine learning is employed to automatically detect such potentially unfair clauses in online contracts. Results show that the proposed system could provide a valuable tool for lawyers and consumers alike.

No abstract available.

The automatic extraction of arguments from text, also known as argument mining, has recently become a hot topic in artificial intelligence. Current research has only focused on linguistic analysis. However, in many domains where communication may be also vocal or visual, paralinguistic features too may contribute to the transmission of the message that arguments intend to convey. For example, in political debates a crucial role is played by speech. The research question we address in this work is whether in such domains one can improve claim detection for argument mining, by employing features from text and speech in combination. To explore this hypothesis, we develop a machine learning classifier and train it on an original dataset based on the 2015 UK political elections debate.

In this paper, we motivate the use of models and algorithms from the area of Statistical Relational Learning (SRL) as a framework for the description and the analysis of games. SRL combines the powerful formalism of first-order logic with the capability of probabilistic graphical models in handling uncertainty in data and representing dependencies between random variables: for this reason, SRL models can be effectively used to represent several categories of games, including games with partial information, graphical games and stochastic games. Inference algorithms can be used to approach the opponent modeling problem, as well as to find Nash equilibria or Pareto optimal solutions. Structure learning algorithms can be applied, in order to automatically extract probabilistic logic clauses describing the strategies of an opponent with a high-level, human-interpretable formalism. Experiments conducted using Markov logic networks, one of the most used SRL frameworks, show the potential of the approach.

Argumentation mining aims at automatically extracting structured arguments from unstructured textual documents. It has recently become a hot topic also due to its potential in processing information originating from the Web, and in particular from social media, in innovative ways. Recent advances in machine learning methods promise to enable breakthrough applications to social and economic sciences, policy making, and information technology: something that only a few years ago was unthinkable. In this survey article, we introduce argumentation models and methods, review existing systems and applications, and discuss challenges and perspectives of this exciting new research area.

The literature on short-term traffic flow forecasting has undergone great development recently. Many works, describing a wide variety of different approaches, which very often share similar features and ideas, have been published. However, publications presenting new prediction algorithms usually employ different settings, data sets, and performance measurements, making it difficult to infer a clear picture of the advantages and limitations of each model. The aim of this paper is twofold. First, we review existing approaches to short-term traffic flow forecasting methods under the common view of probabilistic graphical models, presenting an extensive experimental comparison, which proposes a common baseline for their performance analysis and provides the infrastructure to operate on a publicly available data set. Second, we present two new support vector regression models, which are specifically devised to benefit from typical traffic flow seasonality and are shown to represent an interesting compromise between prediction accuracy and computational efficiency. The SARIMA model coupled with a Kalman filter is the most accurate model; however, the proposed seasonal support vector regressor turns out to be highly competitive when performing forecasts during the most congested periods.

The Single Frontier Bi-Directional Search (SBS) framework was recently introduced. A node in SBS corresponds to a pair of states, one from each of the frontiers and it uses front-to- front heuristics. In this paper we present an enhanced version of SBS, called eSBS, where pruning and caching techniques are applied, which significantly reduce both time and memory needs of SBS. We then present a hybrid of eSBS and IDA∗ which potentially uses only the square root of the memory required by A∗ but enables to prune many nodes that IDA∗ would generate. Experimental results show the benefit of our new approaches on a number of domains.

Prediction of binding sites from sequence can significantly help toward determining the function of uncharacterized proteins on a genomic scale. The task is highly challenging due to the enormous amount of alternative candidate configurations. Previous research has only considered this prediction problem starting from 3D information. When starting from sequence alone, only methods that predict the bonding state of selected residues are available. The sole exception consists of pattern-based approaches, which rely on very specific motifs and cannot be applied to discover truly novel sites. We develop new algorithmic ideas based on structured-output learning for determining transition-metal-binding sites coordinated by cysteines and histidines. The inference step (retrieving the best scoring output) is intractable for general output types (i.e., general graphs). However, under the assumption that no residue can coordinate more than one metal ion, we prove that metal binding has the algebraic structure of a matroid, allowing us to employ a very efficient greedy algorithm. We test our predictor in a highly stringent setting where the training set consists of protein chains belonging to SCOP folds different from the ones used for accuracy estimation. In this setting, our predictor achieves 56 percent precision and 60 percent recall in the identification of ligand-ion bonds.

MOTIVATION: Accurate prediction of contacts between beta-strand residues can significantly contribute towards ab initio prediction of the 3D structure of many proteins. Contacts in the same protein are highly interdependent. Therefore, significant improvements can be expected by applying statistical relational learners that overcome the usual machine learning assumption that examples are independent and identically distributed. Furthermore, the dependencies among beta-residue contacts are subject to strong regularities, many of which are known a priori. In this article, we take advantage of Markov logic, a statistical relational learning framework that is able to capture dependencies between contacts, and constrain the solution according to domain knowledge expressed by means of weighted rules in a logical language. RESULTS: We introduce a novel hybrid architecture based on neural and Markov logic networks with grounding-specific weights. On a non-redundant dataset, our method achieves 44.9% F(1) measure, with 47.3% precision and 42.7% recall, which is significantly better (P < 0.01) than previously reported performance obtained by 2D recursive neural networks. Our approach also significantly improves the number of chains for which beta-strands are nearly perfectly paired (36% of the chains are predicted with F(1) >or= 70% on coarse map). It also outperforms more general contact predictors on recent CASP 2008 targets.

UNLABELLED: The web server MetalDetector classifies histidine residues in proteins into one of two states (free or metal bound) and cysteines into one of three states (free, metal bound or disulfide bridged). A decision tree integrates predictions from two previously developed methods (DISULFIND and Metal Ligand Predictor). Cross-validated performance assessment indicates that our server predicts disulfide bonding state at 88.6% precision and 85.1% recall, while it identifies cysteines and histidines in transition metal-binding sites at 79.9% precision and 76.8% recall, and at 60.8% precision and 40.7% recall, respectively. AVAILABILITY: Freely available at http://metaldetector.dsi.unifi.it. SUPPLEMENTARY INFORMATION: Details and data can be found at http://metaldetector.dsi.unifi.it/help.php.