Nowadays, the spread of deceptive reviews is a problem that has reached critical dimensions, having a significant economic impact on business activities. This paper aims to estimate - at the quantitative and qualitative levels - the possibility of using particular words to disambiguate between truthful and deceptive text, focusing on reviews produced in the cultural heritage domain. For this purpose, a lexicon-based methodology has used two different lexicons: sentiment information, intensifiers, downtoners, and negation operators. As known in the literature, these elements are crucial in a classification process related to deceptiveness. The evaluation phase has considered quantitative metrics such as accuracy and F1 score and ad hoc developed metrics that consider specific linguistic parameters such as polarity and tone of voice intensifiers. A qualitative analysis of a subset of the corpus has also been carried out to understand better factors that impact the classification of deceptive review. Several linguistic features have been considered, ranging from the number of intensifiers to their type and position in phrases and sentences. A comparison between the performances of two different lexicons used has been added to the analysis.

Anomaly detection in industrial environments aims at detecting anomalies in the monitoring data of industrial machinery or equipment, as soon as possible, preferably presenting real-time alarms, to alert the monitoring staff and start maintenance activities timely. In this paper, the problem of anomaly detection of an industrial furnace is tackled, for the real-time recognition of punctual anomalies on multivariate time series. To this aim, a real-time anomaly detection approach is proposed: first, time series acquired from the real machinery are filtered, to select those of interest for possible anomalies, and pre-processed, to obtain sliding windows for real-time detection, then distinct univariate models are applied, to identify different anomaly types. For the application considered here, data regarding the machinery behaviour were available only for normal functioning, thus an unsupervised approach is chosen. In particular, deep learning models based on autoencoders are used to detect punctual anomalies, by reconstructing each window and evaluating the reconstruction error of its last point. An extensive set of autoencoder models is proposed, with varying architecture in terms of type of model (vanilla/variational autoencoders), type of layers (fully connected/LSTM/BiLSTM), and hyperparameters (number of layers, intermediate sizes, BiLSTM type). Available data are split, and used to train the models, and to test them on the normal signal and on synthetic anomalies injected on it, which are of particular interest and were designed according to domain experts. Performances of the proposed models show differences among them, depending on the model architecture. The most efficient models, in terms of F1 score of detection and number of parameters, are identified by their t-test comparison, and the capability of detecting anomalies online is demonstrated. In particular, the proposed anomaly detection approach, including a selected autoencoder with LSTM layers, is able to correctly recognize normal trends, with very few false positives, and promptly give alarms as different anomalous trends appear.

Abstract In the last decade, the demand for readily accessible corpora has touched all areas of natural language processing, including coreference resolution. However, it is one of the least considered sub-fields in recent developments. Moreover, almost all existing resources are only available for the English language. To overcome this lack, this work proposes a methodology to create a corpus for coreference resolution in Italian exploiting knowledge of annotated resources in other languages. Starting from OntonNotes, the methodology translates and refines English utterances to obtain utterances respecting Italian grammar, dealing with language-specific phenomena and preserving coreference and mentions. A quantitative and qualitative evaluation is performed to assess the well-formedness of generated utterances, considering readability, grammaticality, and acceptability indexes. The results have confirmed the effectiveness of the methodology in generating a good dataset for coreference resolution starting from an existing one. The goodness of the dataset is also assessed by training a coreference resolution model based on BERT language model, achieving the promising results. Even if the methodology has been tailored for English and Italian languages, it has a general basis easily extendable to other languages, adapting a small number of language-dependent rules to generalize most of the linguistic phenomena of the language under examination.

During the last years, the significant increase in the number of patients in need of rehabilitation has generated an unsustainable economic impact on healthcare systems, implying a reduction in therapeutic supervision and support for each patient. To address this problem, this paper proposes a telerehabilitation system based on serious games and in-cloud data analytics services, in accordance with Industry 4.0 design principles regarding modularity, service orientation, decentralization, virtualization, and real-time capability. The system, specialized for poststroke patients, comprises components for real-time acquisition of patient's motor data and a decision support service for their analysis. Raw data, reports, and recommendations are made available on the cloud to clinical operators to remotely assess rehabilitation outcomes and dynamically improve therapies. Furthermore, the results of a pilot study on the clinical impact deriving from the adoption of the proposed solution, and of a qualitative analysis about its acceptance, are presented and discussed.

Question Classification is a core module of Question Answering paradigm. Development of classification models based on neural networks showed that convolutional architectures allow obtaining uppermost results for this task. In particular, this type of approach avoids extracting features from questions, by treating text as a sequence of words, and transforming each word in a dense vector, named word embedding. Among different techniques to learn word embeddings, a recent approach takes into account also subword information, which could be very useful for morphologically rich languages. In this paper, a Question Classification approach based on word embedding using subword information and Convolutional Neural Networks is proposed, in order to improve classification accuracy. In particular, questions taken from a TRC dataset are considered, and a comparison between English and Italian languages is reported, by highlighting eventual improvements obtained by initializing word embeddings with advanced vectors learned in an unsupervised manner using skip- gram model and comprising character-based information.

Question answering (QA) is a challenging task and has received considerable attention in the last years. Answer selection among candidate answers is one of the main phases for QA and the best answer to be returned is determined in this phase. A common approach consists in considering the selection of the final answer(s) as a ranking problem. So far, different methods have been proposed, mainly oriented to produce a single best ranking model operating in the same way on all the question types. Differently, this paper proposes a fuzzy approach for ranking and selecting the correct answer among a list of candidates in a state-of-the-art QA system operating with factoid and description questions on Italian corpora pertaining a closed domain. Starting from the consideration that this ranking problem can be reduced to a classification one, the proposed approach is based on the Likelihood-Fuzzy Analysis (LFA), applied in this case for mining fuzzy rule-based models able to discern correct (True) from incorrect answers (False). Such fuzzy models are mined as specifically tailored to each question type, and, thus, can be individually applied to produce a more robust and accurate final ranking. An experimental session over a collection of questions pertaining the Cultural Heritage domain, using a manually annotated gold-standard dataset, shows that considering specific fuzzy ranking models for each question type improves the accuracy of the best answer returned back to the user.

In the last years, the interest about enhancing the interface usability of applications has strongly increased, focusing, in particular, on chatbots, i.e. conversational agent that interacts with users, turn by turn using natural language. However, building chatbots for answering to questions over structured medical knowledge bases is a very thorny task and is still considered an open research challenge. In order to face this issue, this paper proposes a knowledge-based conversational chatbot for medical question answering, aimed at supporting: i) the formulation of factoid questions over medical knowledge bases; ii) the generation of more precise and contextualized dialog responses by analyzing the relations between entities in knowledge bases; iii) the detection of ambiguous user intents, with respect to the current dialog state and the suggestion of some interaction hints aimed at clarifying and/or confirming their meaning. A relevant characteristic of this system is represented by the usage of Knowledge Graphs to formally represent textual inputs given by the user as well as templates of questions and, contextually, efficiently navigate and use the domain knowledge of interest to provide an answer. The proposed chatbot has been implemented as a desktop application named “Medical Assistant” able to conversate with users interested to diagnose and identify the possible diseases causing a symptom, and find the most suitable treatment for a medical problem. It has been proficiently tested with respect to some factoid questions, showing its capability to help user reach the desired information also in the case of initial missing information.

Question Classification (QC) is a very important module, to include into the pipeline usually employed to implement the Question Answering paradigm. Recently, good results have been achieved on the QC task by using Convolutional Neural Networks (CNNs). This approach requires setting a CNN architecture and a huge number of hyperparameters to obtain the desirable achievements, and only little research has been addressed on this activity. Moreover, while the greatest part of research strength focused on English language, very few works dealt with other languages. In this work, an approach based on neural networks is used to classify Italian questions taken from a TREC dataset. In particular, different solutions regarding the CNN architecture are tested, and, according to literature advices, the best settings are searched in the proper ranges, in order to maximize the classification power for the particular case of Italian questions dataset.

Classification systems based on Fuzzy Logic are of particular importance in the ambit of cognitive systems, due to their ability of managing uncertainty and presenting interpretable knowledge bases by emulating human cognition processes. However, the notion of interpretability is not yet exhaustively defined. In this work, the properties assessing the interpretability of a fuzzy classifier are discussed, and on this basis two indexes are proposed, which numerically evaluate readability and semantic interpretability, respectively. These indexes are calculated for a benchmark dataset, showing how they can be used to quantitatively compare systems with different characteristics.

Recently, innovative and mobile health services have been developed by embedding knowledge-based systems, with the aim of remotely promoting wellness and healthy lifestyle, monitoring patients' chronic diseases and improving their adherence to therapies. Even if different knowledge-based systems have been proposed for mobile devices, they are typically based on precise production rules built on the top of ontological primitives for describing the domain of interest. Thus, they are not able to handle medical knowledge graded and affected by uncertainty, which often underlies medical decision-making processes. In order to address this topic, this paper presents a hybrid, rule-based reasoning system for mobile devices aimed at enabling the realization of intelligent health services. This system is essentially characterized by two main features: i) a hybrid knowledge representation approach for modelling productions rules involving both precise and vague information by integrating ontological and fuzzy primitives; ii) a lazy reasoning algorithm able to efficiently process this hybrid knowledge and timely produce answers. A case study has been arranged in order to evaluate the effectiveness of the proposed system within a mobile application for detecting heart arrhythmias.

Technological innovations coupled with the rapidly expanding amount of medical data digitally collected and stored have made possible to develop advanced Decision Support Systems (DSSs) able to aid physicians in medical diagnosis, by helping them in classifying among different diseases. Building this typology of DSSs preliminarily requires the extraction from huge sets of clinical data of hidden relationships between possible classes and known features. This task is very thorny due to information uncertainties affecting features as well as biases regarding prior probabilities of classes, depending on the environment and conditions of data collection. To address these problems, this paper presents an approach for building an interval type-2 fuzzy DSS able to handle information uncertainties and perform unbiased medical diagnoses, by adapting interval type-2 fuzzy sets to a medical dataset. A proof of concept is given by applying the proposed approach on a benchmark medical dataset. A comparison is performed between i) a type-1 fuzzy system with prior probabilities extracted from the dataset; ii) an interval type-2 fuzzy DSSs modelling non-biased prior probabilities; and iii) the best known classification methods. This comparison is made both when prior probabilities are fixed, i.e. when the system is evaluated by a stratified cross-validation technique, or while no assumption is made about prior probabilities. Results show that the proposed approach reach good performance in most of the cases, thus evidencing the usefulness of extracting an interval type-2 fuzzy system not based on biased prior probabilities, for obtaining more reliable results.

Clinical Decision Support Systems (CDSSs) are typically based on clinical guidelines explicitly formalized in the form of rules for reproducing the physician's decisionmaking process and, also, improving the efficiency of medical practices. With the aim of building CDSSs able to represent uncertainty existing in clinical guidelines and efficiently reason on a huge number of inter-connected rules, this paper presents a multi-level fuzzy inference system offering the following set of specifically-devised functionalities: (i) fuzzy rules can be organized in one or more groups of positive evidence rules, where each group is able to interact with other ones by properly chaining their conclusions; (ii) rules inside a group are independently processed and evaluated; (iii) each group of rules can be customized by means of a peculiar configuration for the inference; (iv) a fuzzy ELSE rule can be associated to a group for assembling all the negative evidence for a specific situation. A proof of concept scenario is finally given to describe how the proposed solution can be applied.

BACKGROUND: The diagnosis of many diseases can be often formulated as a decision problem; uncertainty affects these problems so that many computerized Diagnostic Decision Support Systems (in the following, DDSSs) have been developed to aid the physician in interpreting clinical data and thus to improve the quality of the whole process. Fuzzy logic, a well established attempt at the formalization and mechanization of human capabilities in reasoning and deciding with noisy information, can be profitably used. Recently, we informally proposed a general methodology to automatically build DDSSs on the top of fuzzy knowledge extracted from data. METHODS: We carefully refine and formalize our methodology that includes six stages, where the first three stages work with crisp rules, whereas the last three ones are employed on fuzzy models. Its strength relies on its generality and modularity since it supports the integration of alternative techniques in each of its stages. RESULTS: The methodology is designed and implemented in the form of a modular and portable software architecture according to a component-based approach. The architecture is deeply described and a summary inspection of the main components in terms of UML diagrams is outlined as well. A first implementation of the architecture has been then realized in Java following the object-oriented paradigm and used to instantiate a DDSS example aimed at accurately diagnosing breast masses as a proof of concept. CONCLUSIONS: The results prove the feasibility of the whole methodology implemented in terms of the architecture proposed.

The formalization and manipulation of complex and not yet assessed rules by clinicians are critical for Decision Support Systems (DSSs) performance in supporting remote monitoring of chronic patients. Sometimes, structural anomalies, such as inconsistency and redundancy, can occur. This work presents a novel system, named Consistency Checker, aimed at verifying the reliability of conditionaction clinical rules in knowledge-based DSSs. This system allows the verification of very complex rules having in their antecedent parts not only simple logical conditions, but also arithmetical expressions. Moreover, the Consistency Checker provides a new classification of the detected anomalies, aimed at fully describing the rule verification results, as well as a suitable knowledge representation formalism to encode condition-action rules in a general way. The system has been designed according to the Service Oriented Architecture and implemented as a Web Service within the CHRONIOUS Project.

Fuzzy-based Decision Support Systems (DSSs) have gained increasing importance in medicine, since they rely on a transparent and interpretable rule base. A very attractive feature for these systems is to present their results as a set of plausible conclusions, each of them associated with a degree of possibility. In order to face this need, this work proposes a novel approach consisting in hybridization of possibility theory and a classical fuzzy clustering method, based on a distance metric interpretable in a probabilistic framework with the final aim of determining both fuzzy rules and partitions. As a proof of concept, the method has been applied to a real-life application pertaining the classification of Multiple Sclerosis Lesions. Finally, some sophistications are proposed for a future refinement, in order to improve the quality of results and the generality of applications.

Remote Health Monitoring (RHM) programmes are being increasingly developed to face the pervasive diffusion of chronic diseases. RHM strongly relies on ICT intelligent platforms devised to remotely acquire multisource data, process these according to specific domain knowledge and support clinical decision making. However, since RHM domain is continuously evolving and the pertinent knowledge is not yet consolidated, there is a great demand for services and tools that allow the encoded knowledge to be modified and enriched. This paper presents a Knowledge Editing Service (KES), which aims at enabling clinicians to insert novel knowledge, in a controlled fashion, into an ICT intelligent platform. The solution proposed is innovative since it addresses synergistically peculiar issues related to (i) RHM knowledge format; (ii) controlled editing patterns; (iii) knowledge verification and (iv) cooperative knowledge editing. None of the existing methods and systems for knowledge authoring tackles all these aspects at the same time. A prototype of the KES has been implemented and evaluated in real operational conditions.

Recent research agrees on the utility of fuzzy reasoning for the development of Decision Support Systems, which help to classify clinical data. In this context, methods or techniques for representing fuzzy terms in the form of interpretable fuzzy sets obtained from numerical data are strongly required. Typically, in medical settings, statistical data are available or can be obtained from rough data, in the form of probability distributions or likelihood functions. Until now, no theoretical approach was proposed for transforming a probability distribution into a likelihood view fuzzy set. In this paper, a method is developed which generalizes some existing approaches by giving them a theoretical justification. The method enables the construction of normal fuzzy sets, which can be chosen to have a triangular or trapezoidal shape where lateral edges are adapted depending on the input probability distribution. The method was assessed through its application to a simulated normal probability distribution and to real case study pertaining the classification of Multiple Sclerosis lesions.

Pervasive Computing technologies can be successfully used in hospital and clinical environments in order to provide smart and enhanced healthcare services and applications. According to this vision, the paper presents an RFID-based application for handling the workflow of patients injected with radi

The advancements of wireless and mobile computing technologies and the diffusion of pervasive healthcare technologies are changing our perception of healthcare. In this paper, we describe how the pervasive computing technologies can be used to build a Smart Hospital. In particular, we propose a concrete implementation of a smart hospital and discuss how e-Health services and applications can be enhanced by location information. As a solution, we present semantic models, mechanisms and a system to locate diverse kinds of mobile entities in Smart Hospitals. The key feature of the system is the semantic integration of different positioning systems that not only enables the hospital to transparently handle such physical positioning systems, but also to reason on location information coming from different systems and to combine them in order to get higher context information or to resolve inconsistencies or conflicts due to sensing errors or limitations of the positioning systems.