Ciro Castiello

• University of Bari Aldo Moro

In Explainable Artificial Intelligence, the interpretation of the decisions provided by a model is of primary importance. In this context, we consider Fuzzy C-Means (FCM), which is a clustering algorithm that induces a model from data by assigning, to each data-point, a degree of membership to each cluster such that the sum of memberships is one. A...

Image segmentation is the process of dividing a digital image into multiple segments or regions, each of which represents a different object or part of the image or shares certain visual characteristics. The goal of image segmentation is to simplify the representation of an image into something that is more meaningful and easier to analyze. Typical...

We study the impact of fuzziness on the behavior of Fuzzy Rule-Based Classifiers (FRBCs) defined by trapezoidal fuzzy sets forming Strong Fuzzy Partitions. In particular, if an FRBC selects the class related to the rule with the highest activation (so-called Winner-Takes-All approach), then fuzziness, as quantified by the slope of the membership fu...

The semantic segmentation of remotely sensed images is a difficult task because the images do not represent well-defined objects. To tackle this task, fuzzy logic represents a valid alternative to convolutional neural networks-especially in the presence of very limited data-, as it allows to classify these objects with a degree of uncertainty. Unfo...

Current UAV technology has led to an exponential growth in the potential applications of drones in both the military and the civilian fields. Therefore, drone missions must be subject to regulations that ensure safety in operations, especially in Urban Air Mobility tasks. Nevertheless, unpredictable conditions may lead drones to dangerous routes, e...

Crowd counting on the drone platform is an interesting topic in computer vision, which brings new challenges such as small object inference, background clutter and wide viewpoint. However, there are few algorithms focusing on crowd counting on the drone-captured data due to the lack of comprehensive datasets. To this end, we collect a large-scale d...

Since 2016, there is an increasing interest on research topics such as fairness, accountability and interpretability, in the entire community of researchers in Artificial Intelligence. However, there were researchers working hard on these research topics much earlier. For example, Michalski published his comprehensibility postulate in the 1980s whi...

Fuzzy sets and fuzzy logic are powerful tools widely used to represent human knowledge and mimic human reasoning capabilities, being the main constituents of fuzzy systems. Among the different approaches to fuzzy systems, fuzzy rule-based systems represent the one offering a better framework for interpretability considerations. Their applications r...

Interpretability is one of the most valuable properties of fuzzy systems. Despite the effort made by the research community for characterizing interpretability, there is not a consensus about how to measure interpretability yet. It is admitted that the analysis of interpretability is subjective because it depends on the background of the person who...

Fuzzy systems have found widespread application in several contexts and proved their suitability in tackling a number of diverse real-world problems. However, the realization of such systems must be well grounded on some solid theoretical bases that scientists and developers should properly master. In this chapter we discuss the key elements of fuz...

Explainable Artificial Intelligence is a novel paradigm conjugating the effectiveness of machine learning with the new requirements coming from the integration of intelligent systems in the human society. Explainable Artificial Intelligence can find successful application in a plethora of contexts, endowing classical intelligent systems with a cruc...

We describe step by step how to design, implement and validate an interpretable fuzzy rule-based beer style classifier endowed with explanation capability. First, we revise some preliminary work regarding both interpretable fuzzy modeling methodologies and related software. Second, we introduce the use case on beer style classification. Third, we b...

Fuzzy systems are commonly considered suitable tools to express knowledge in a human comprehensible fashion. This kind of characterization makes them eligible for being applied in several contexts where interpretability is a major issue and humans may profit from a self-explanatory form of automatic computation. However, fuzzy systems are not inter...

The importance of Trustworthy and Explainable Artificial Intelligence (XAI) is recognized in academia, industry and society. This book introduces tools for dealing with imprecision and uncertainty in XAI applications where explanations are demanded, mainly in natural language. Design of Explainable Fuzzy Systems (EXFS) is rooted in Interpretable F...

All details about the session are at: https://sites.google.com/view/xai-fuzzieee2021 The aim of this session is to offer an opportunity for researchers and practitioners to identify new promising research directions on eXplainable Artificial Intelligence (XAI) and to provide a forum to disseminate and discuss XAI, with special attention to Interpr...

Unmanned aerial vehicles (UAVs), more commonly known as drones, are increasingly used as a technological support tool for search-and-rescue (SAR) operations and also for post-disaster area explorations. UAVs can be equipped with high-resolution cameras and embed GPUs powerful enough to provide effective and efficient aid to emergency rescue operati...

We study the influence of fuzziness of trapezoidal fuzzy sets in the strong fuzzy partitions (SFPs) that constitute the database of a fuzzy rule-based classifier. To this end, we develop a particular representation of the trapezoidal fuzzy sets that is based on the concept of cuts, which are the cross-points of fuzzy sets in a SFP and fix the posit...

Crowd counting on the drone platform is an interesting topic in computer vision, which brings new challenges such as small object inference, background clutter and wide viewpoint. However, there are few algorithms focusing on crowd counting on the drone-captured data due to the lack of comprehensive datasets. To this end, we collect a large-scale d...

This paper proposes a novel lightweight and fast convolutional neural network to learn a regression model for crowd counting in images captured from drones. The learning system is initially based on a multi-input model trained on two different views of the same input for the task at hand: (i) real-world images; and (ii) corresponding synthetically...

Crowd analysis is receiving an increasing attention in the last years because of its social and public safety implications. One of the building blocks of crowd analysis is crowd counting and the associated crowd density estimation. Several commercially available drones are equipped with on-board cameras and embed powerful GPUs, making them an excel...

This is a classification dataset made up of 400 instances perfectly balanced with 50 instances per class. The classification task consists of identifying one out of 8 beer styles (Blanche, Lager, Pilsner, IPA, Stout, Barleywine, Porter, and Belgian Strong Ale) in terms of 3 attributes (color, bitterness and strength). The original file is available...

Unmanned aerial vehicles (UAVs) also known as drones are increasingly populating our skies. This represents a relevant issue both for the legislator and the researcher. While the regulation plans often assume precautionary approaches, stating restrictive conditions of use for the sake of public safety, the applied research is exploring novel strate...

In this paper, we propose a novel crowd detection method for drone safe landing, based on an extremely light and fast fully convolutional neural network. Such a computer vision application takes advantage of the technical tools some commercial drones are equipped with. The proposed architecture is based on a two-loss model in which the main classif...

JFML is an open source Java library aimed at facilitating interoperability of fuzzy systems by implementing the IEEE Std 1855-2016-the IEEE Standard for Fuzzy Markup Language (FML) that is sponsored by the IEEE Computational Intelligence Society. We developed a Python wrapper for JFML that enables to use all the functionalities of JFML through a Py...

In recent years, there has been a huge effort connecting all kind of devices to Internet. From small devices (e.g., e-health monitoring sensors or mobile phones) that we carry daily in what is called the body-area-network, to big devices (such as cars), passing by all devices (e.g., TVs or refrigerators) at home. In modern cities, everything (at wo...

In medical problems both the information and the reasoning used by clinicians for drawing conclusions about patients’ health are inherently uncertain and vague. Fuzzy logic is a powerful tool for representing and handling this uncertainty, leading to fuzzy systems that can support decisions in medical diagnosis. In this work we propose a fuzzy rule...

This is the call for paper of a session that is proposed to be held in FUZZ-IEEE 2019

Fuzzy rule-based systems are effective tools for acquiring knowledge from data and represent it in a linguistically interpretable form. To achieve interpretability, input features are granulated in fuzzy partitions. A critical design decision is the selection of the granularity level for each input feature. This paper presents an approach, called D...

Purpose: In this paper we propose a framework for intelligent analysis of Twitter data. The purpose of the framework is to allow users to explore a collection of tweets by extracting topics with semantic relevance. In this way, it is possible to detect groups of tweets related to new technologies, events and other topics that are automatically dis...

This paper describes how to build an eXplainable Artificial Intelligence (XAI) classifier for a real use case related to beer style classification. It combines an opaque machine learning algorithm (Random Forest) with an interpretable machine learning algorithm (Decision Tree). The result is a XAI classifier which provides users with a good interpr...

The IV European Summer School on Fuzzy Logic and Applications is promoted by the European Society for Fuzzy Logic and Technology. PhD students and young researchers represent the ideal audience for the School which aims at introducing the core aspects and the recent developments of Fuzzy Logic and related applications. The School proposes severa...

This paper presents a novel hybrid approach for building eXplainable Artificial Intelligence (XAI) systems. It combines an opaque machine learning system with several interpretable systems to build a whole XAI system, i.e., a system which provides users with a good interpretability-accuracy trade-off but also with explanation capabilities. First, t...

The goal of this special session is to discuss and disseminate the most recent advancements focused on explainable artificial intelligence. The session goes a step ahead with respect to the previous events we organized (which were mainly focused on interpretable fuzzy systems) in some other conferences: joint IFSA-EUSFLAT 2009, ISDA 2009, WCCI 2010...

This paper presents the results of a bibliometric study of the recent research on eXplainable Artificial Intelligence (XAI) systems. We took a global look at the contributions of scholars in XAI as well as in the subfields of AI that are mostly involved in the development of XAI systems. It is worthy to remark that we found out that about one third...

In modern cities, everything is connected to Internet and the amount of data available on-line grows dramatically. Humans face two main challenges: i) to extract valuable knowledge from the Big Data; ii) to become part of the equation as active actors in the Internet of Things. Fuzzy intelligent systems are currently used in many applications in th...

This session goes on with the previous events (focus on interpretable fuzzy systems) that we already organized. Further info at http://tiny.cc/ifs-fuzzieee2017

In this paper we face the problem of intelligently analyze Twitter data. We propose a novel workflow based on Nonnegative Matrix Factorization (NMF) to collect, organize and analyze Twitter data. The proposed workflow firstly fetches tweets from Twitter (according to some search criteria) and processes them using text mining techniques; then it is...

In this paper we illustrate the use of Nonnegative Matrix Factorization (NMF) to analyze real data derived from an e-learning context. NMF is a matrix decomposition method which extracts latent information from data in such a way that it can be easily interpreted by humans. Particularly, the NMF of a score matrix can automatically generate the so c...

This paper introduces FISDeT, a tool to support the design of Fuzzy Inference Systems, composed of a set of Python modules sharing the standard specification language FCL used for FIS definition. FISDeT includes a graphical user interface that enables easy definition and quick update of elements composing the knowledge base of a FIS. Given the know...

Structured Abstract Purpose – A key task in many Collective Intelligence systems is to represent people and resources in some computational form that is general enough to accommodate different needs and diverse sources of information. Also, it should take into account the inevitable imprecision deriving from the necessity of representing complex ph...

Many fuzzy algorithms and models are indeed aimed at extracting knowledge from data, and the acquired knowledge must be usually communicated to users. However, as far as such knowledge is difficult to understand by users, the acceptance of such methods may be seriously compromised. Interpretability must be the central point on fuzzy system modeling...

Fuzzy systems are universally acknowledged as valuable tools to model complex phenomena while preserving a readable form of knowledge representation. The resort to natural language for expressing the terms involved in fuzzy rules, in fact, is a key factor to conjugate mathematical formalism and logical inference with human-centered interpretability...

DC* (double clustering with A*) is an algorithm capable of generating highly interpretable fuzzy information granules from pre-classified data. These information granules can be used as bulding-blocks for fuzzy rule-based classifiers that exhibit a good tradeoff between interpretability and accuracy. DC* relies on A* for the granulation process, wh...

The adoption of triangular fuzzy sets to define Strong Fuzzy Partitions (SFPs) is a common practice in the research community: due to their inherent simplicity, triangular fuzzy sets can be easily derived from data by applying suitable clustering algorithms. However, the choice of triangular fuzzy sets may be limiting for the modeling process. In t...

In this paper we compare two algorithms that are capable of generating fuzzy partitions from data so as to verify a number of interpretability constraints: Hierarchical Fuzzy Partitioning (HFP) and Double Clustering with A* (DC*). Both algorithms exhibit the distinguishing feature of self-determining the number of fuzzy sets in each fuzzy partition...

Decision support systems in Medicine must be easily comprehensible, both for physicians and patients. In this chapter, the authors describe how the fuzzy modeling methodology called HILK (Highly Interpretable Linguistic Knowledge) can be applied for building highly interpretable fuzzy rule-based classifiers (FRBCs) able to provide medical decision...

In questo articolo proponiamo l’impiego delle fattorizzazioni matriciali non negative per l’analisi dei dati nell’Educational Data Mining. Il metodo si basa su un processo di decomposizione di un dataset per l’estrazione di informazioni latenti di immediata interpretazione. In particolare, l’applicazione delle fattorizzazioni non negative a score m...

Decision support systems in Medicine must be easily comprehensible, both for physicians and patients. In this chapter, the authors describe how the fuzzy modeling methodology called HILK (Highly Interpretable Linguistic Knowledge) can be applied for building highly interpretable fuzzy rule-based classifiers (FRBCs) able to provide medical decision...

When approaching real-world problems with in- telligent systems, an interaction with user is often expected. However, data-driven models are usually evaluated only in terms of accuracy, thus not involving users. In literature several works have been proposed for defining measures for interpretability assessment, however, such measures are mostly ba...

In computing with words (CWW), knowledge is linguistically represented and has an explicit semantics defined through fuzzy information granules. The linguistic representation, in turn, naturally bears an implicit semantics that belongs to users reading the knowledge base; hence a necessary condition for achieving interpretability requires that impl...

The aim of the work is to show the potential usefulness of interpretable fuzzy modeling for decision support in medical applications. For this pursuit, we present an approach for designing interpretable fuzzy systems concerning the prognosis prediction in Immunoglobulin A Nephropathy (IgAN). To deal with such a hard problem, prognosis has been gran...

The common practices of machine learning appear to be frustrated by a number of theoretical results denying the possibility of any meaningful implementation of a “superior” learning algorithm. However, there exist some general assumptions that, even when overlooked, preside the activity of researchers and practitioners. A thorough reflection over s...

Computing with words (CWW) relies on linguistic representation of knowledge that is processed by operating at the semantical level defined through fuzzy sets. Linguistic representation of knowledge is a major issue when fuzzy rule based models are acquired from data by some form of empirical learning. Indeed, these models are often requested to exh...

A key feature for machine intelligence is the ability of learning knowledge from past experiences. Furthermore, in a human-centric environment, the acquired knowledge must fulfill comprehensibility requirements so as to be shared by human users. In literature, several approaches have been proposed to acquire comprehensible knowledge from data by pr...

Recommender systems are systems capable of assisting users by quickly providing them with relevant resources according to their interests or preferences. The efficacy of a recommender system is strictly connected with the possibility of creating meaningful user profiles, including information about user preferences, interests, goals, usage data and...

Adaptive software systems are systems that tailor their behavior to each user on the basis of a personalization process. The efficacy of this process is strictly connected with the possibility of an automatic detection of preference profiles, through the analysis of the users' behavior during their interactions with the system. The definition of su...

Interpretability represents the most important driving force behind the implementation of fuzzy logic-based systems. It can be directly related to the system's knowledge base, with reference to the human user's easiness experienced while reading and understanding the embedded pieces of information. In this paper, we present a preliminary study on i...

Interpretability is one of the most important driving forces for the adoption of fuzzy rule-based classifiers. However, it is not given for granted, especially when fuzzy models are acquired from data. Therefore, evaluation of interpretability should be regarded as a major research topic. In this paper, we describe a technique for automatic interpr...

In this chapter an analysis of computational mechanisms of induction is brought forward, in order to assess the potentiality of meta-learning methods versus the common base-learning practices. To this aim, firstly a formal investigation of inductive mechanisms is accomplished, sketching a distinction between fixed and dynamical bias learning. Then...

We describe an automatic approach for evaluating in- terpretability of fuzzy rule-based classifiers. The approach is based on the logical view of fuzzy rules, which are interpreted as rows in truth tables. These truth tables are subject of a minimization pro- cedure based on a variant of the Quine-McCluskey algorithm. The minimized truth tables are...

Adaptive e-learning systems are growing in popularity in recent years. These systems can offer personalized learn- ing experiences to learners, by supplying each learner with learning contents that meet his/her specific interests and needs. The efficacy of such systems is strictly related to the possibility of automatically deriving models encoding...

This contribution presents a user profile modelling approach based on fuzzy logic techniques. The proposed approach is conceived to find application in various contexts, with the aim of providing personalised contents to different categories of users. Both contents and users are described by metadata, so a description language is introduced along w...

The research activity described in this paper concerns the personalisation process in e-learning contexts. Particular emphasis is laid on the mechanisms of user profiling and association between user profiles and pedagogical resources. A particular profiling model is proposed where both the pedagogical resources and the user profiles are described...

Common inductive learning strategies offer tools for knowledge acquisition, but possess some inherent limitations due to the use of fixed bias during the learning process. To overcome the limitations of such base-learning approaches, a research trend explores the potentialities of meta-learning, which is oriented to the development of mechanisms ba...

This paper presents a user profile modelling approach based on fuzzy logic techniques. The proposed approach is conceived to find application in the context of e-learning processes, with the aim of providing personalised contents to different categories of users. Several concepts are introduced and formalised within a peculiar mathematical framewor...

Text information extraction represents a fundamental issue in the context of digital image processing. Inside this wide area of research, a number of specific tasks can be identified ranging from text detection to text recognition. In this chapter, we deal with the particular problem of text localisation, which aims at determining the exact locatio...

In this work, we propose a new document page segmentation method, capable of differentiating between text, graphics and background, using a neuro-fuzzy methodology. Our approach is based firstly on the analysis of a set of features extracted from the image, available at different resolution levels. An initial segmentation is obtained by classifying...

Questo lavoro presenta uno studio sulla personalizzazione dei contenuti erogati in modalità e-learning, soffermandosi sulla attività relativa alla profilazione degli utenti e alla associazione tra profili utente e risorse didattiche. In particolare, è proposto un modello di profilazione in cui sia le risorse didattiche sia i profili degli utenti so...

In this paper, a novel method for document page segmentation using Wavelet Packet analysis is proposed. To discriminate between text and non-text regions, the image is represented by means of a wavelet packet analysis tree. Successively a feature image is introduced to synthetize the information related to some nodes selected from the quadtree. The...

In recent years, the use of hybrid Soft Computing methods has shown that in various applications the synergism of several techniques is superior to a single technique. For example, the use of a neural fuzzy system and an evolutionary fuzzy system hybridises the approximate reasoning mechanism of fuzzy systems with the learning capabilities of neura...

In this paper we propose a split and merge texture segmentation method. The presented approach is characterised by the introduction of a novel operator, the Local Fuzzy Pattern for texture discrimination, and the employment of a neuro-fuzzy decision support strategy, which supervises the overall split and merge procedure. The effectiveness of the...

The huge amount of image collections connected to multimedia applications has brought forth several approaches to content-based image retrieval, that means retrieving images based on their visual content instead of textual descriptions. In this paper, we present a system, called VIRMA (Visual Image Retrieval by Shape MAtching), which combines diffe...

Successful retrieval of images from large-scale image collections is one of the current problems in the field of multimedia digital libraries. In this context, the use of textual descriptions to represent and query images may provide poor results, due to subjectivity of descriptions. Conversely, image retrieval based on visual content of images has...

Meta-learning practices concern the dynamical search of the bias presiding over the behaviour of artificial learning systems. In this paper we present an original meta-learning framework, namely the Mindful (Meta INDuctive neuro-FUzzy Learning) system. Mindful is based on a neuro-fuzzy learning strategy providing for the inductive processes applica...

In this paper we propose a meta-classification framework which is able to represent the accumulated experience from a base-learner in form of knowledge-base and to exploit it whenever a new task has to be tackled. Our meta-dassificr represents a particular meta-learning strategy where a single learning algorithm is employed both at base- and meta-l...

For pt. I see ibid., (2005). In this paper we investigate the application feasibilities of hybrid systems in meta-learning contexts with a practical perspective. In particular, we present the MINDFUL system, which represents a particular framework for meta-learning founded on the integration of connectionist paradigms and fuzzy knowledge management...

In this paper, we present an original meta-learning framework, namely the Mindful (Meta INDuctive neuro-FUzzy Learning) system. Mindful is based on a neuro-fuzzy learning strategy providing for the inductive processes applicable both to ordinary base-level tasks and to more general cross-task applications. The results of an ensemble of experimental...

This paper describes a peculiar methodology for identifying and classifying the human body structures depicted in digital images. The methodology is articulated in a number of successive steps: an initial process of motion detection recognises human targets into the examined scene; subsequently, a synthetic representation of the previously detected...

Document image analysis represents one of the most relevant topics in the field of image processing: many research efforts have been devoted to devising automatic strategies for document region classification. In this paper, we present a peculiar strategy to extract numerical features from segmented image regions, and their employment for classific...

Common inductive learning strategies offer the tools for knowledge acquisition, but possess some inherent limitations due to the use of fixed bias during the learning process. To overcome limitations of such base-learning approaches, a novel research trend is oriented to explore the potentialities of meta-learning, which is oriented to the developm...

In this paper a neuro-fuzzy modeling framework is proposed, which is devoted to discover knowledge from data and represent it in the form of fuzzy rules. The core of the framework is a knowledge extraction procedure that is aimed to identify the structure and the parameters of a fuzzy rule base, through a two-phase learning of a neuro-fuzzy network...

The task of document image segmentation is to represent a digital image in a more interpretable form, recognising regions containing text, background and graphics. This paper presents a peculiar strategy for document image segmentation, where a neuro-fuzzy approach is involved. Firstly, image is segmented into text, graphics or background during a...

Meta-learning is a novel paradigm aimed to overcome the limitations of traditional base-learning strategies. In particular, meta-learning strategies address the problem of exploiting past experience to dynamically identify the best bias for each task. In this paper, we present a learning framework, built up on the basis of the neuro-fuzzy integrati...

Inductive learning mechanisms offer the tools for knowledge enlargement, but an analysis of common learning strategies reveals the limitations of base-learning methods. The objective of our research consists in defining a meta-learning framework which brings together a base-learner and a meta-learner with the aim of dynamically selecting a proper b...

This paper describes the application of an AI methodol- ogy to the problem of ash property prediction deriving from combustion processes for electric generation. The employed methodology is based on a neuro-fuzzy frame- work that, starting from the analysis of a set of data, de- rives predictive models in form of fuzzy rule base. A par- ticular neu...

This paper briefly surveys the state of the art of a particular mechanism of learning: induction. We discuss inductive mechanisms, drawing attention to the foundation of generalisation success and its lim- itations. The distinction between base-learning and meta-learning approaches is pointed out in order to better identify the peculiar attributes...

We present a neuro-fuzzy approach for classification of image pixels into three classes: contour, regular or texture points. Exploiting the processing capabilities of a neural network, fuzzy classification rules are derived by learning from data and applied to classify pixels in grey-level images. To derive a proper set of training data, the spatia...

In this paper, we propose a neuro-fuzzy modeling framework to discover fuzzy rules and its application to predict chemical properties of ashes produced by thermo-electric generators. The framework is defined by several sequential steps in order to obtain a good predictive accuracy and the readability of the discovered fuzzy rules. First, a feature...

This paper describes a neuro-fuzzy modeling framework for predicting the properties of ashes originated from combustion processes for electric generation. The prediction problem is tackled by means of a neuro-fuzzy system in which a neural network and a fuzzy system are combined in a fused architecture, so that the structure and the parameters of t...