Search

A number of algorithms in the field of artificial intelligence offer poorly interpretable decisions. To disclose the reasoning behind such algorithms, their output can be explained by means of so-called evidence-based (or factual) explanations. Alternatively, contrastive and counterfactual explanations justify why the output of the algorithms is not any different and how it could be changed, respectively. It is of crucial importance to bridge the gap between theoretical approaches to contrastive and counterfactual explanation and the corresponding computational frameworks. In this work we conduct a systematic literature review which provides readers with a thorough and reproducible analysis of the interdisciplinary research field under study. We first examine theoretical foundations of contrastive and counterfactual accounts of explanation. Then, we report the state-of-the-art computational frameworks for contrastive and counterfactual explanation generation. In addition, we analyze how grounded such frameworks are on the insights from the inspected theoretical approaches. As a result, we highlight a variety of properties of the approaches under study and reveal a number of shortcomings thereof. Moreover, we define a taxonomy regarding both theoretical and practical approaches to contrastive and counterfactual explanation.

Artificial Intelligence (AI) is a first class citizen in the cities of the 21st century. In addition, trust, fairness, accountability, transparency and ethical issues are considered as hot topics regarding AI-based systems under the umbrella of Explainable AI (XAI). In this paper we have conducted an experimental study with 15 datasets to validate the feasibility of using a pool of gray-box classifiers (i.e., decision trees and fuzzy rule-based classifiers) to automatically explain a black-box classifier (i.e., Random Forest). Reported results validate our approach. They confirm the complementarity and diversity among the gray-box classifiers under study, which are able to provide users with plausible multi-modal explanations of the considered black-box classifier for all given datasets.

Data-driven classification algorithms have proven highly effective in a range of complex tasks. However, their output is sometimes questioned, as the reasoning behind it may remain unclear due to a high number of poorly interpretable parameters used during training. Evidence-based (factual) explanations for single classifications answer the question why a particular class is selected in terms of the given observations. On the contrary, counterfactual explanations pay attention to why the rest of classes are not selected. Accordingly, we hypothesize that providing classifiers with a combination of both factual and counterfactual explanations is likely to make them more trustworthy. In order to investigate how such explanations can be produced, we introduce a new method to generate factual and counterfactual explanations for the output of pretrained decision trees and fuzzy rule-based classifiers. Experimental results show that unification of factual and counterfactual explanations under the paradigm of fuzzy inference systems proves promising for explaining the reasoning of classification algorithms.

Fairness, Accountability, Transparency and Explainability have become strong requirements in most practical applications of Artificial Intelligence (AI). Fuzzy sets and systems are recognized world-wide because of their outstanding contribution to model AI systems with a good interpretability-accuracy trade-off. Accordingly, fuzzy sets and systems are at the core of the so-called Explainable AI. ExpliClas is a software as a service which paves the way for interpretable and self-explainable intelligent systems. Namely, this software provides users with both graphical visualizations and textual explanations associated with intelligent classifiers automatically learned from data. This paper presents the new functionality of ExpliClas regarding the generation, evaluation and explanation of fuzzy decision trees along with fuzzy inference-grams. This new functionality is validated with two well-known classification datasets (i.e., Wine and Pima), but also with a real-world beer-style classifier.

Diabetes is a serious chronic disease. The importance of clinical decision support systems (CDSSs) to diagnose diabetes has led to extensive research efforts to improve the accuracy, applicability, interpretability, and interoperability of these systems. However, this problem continues to require optimization. Fuzzy rule-based systems are suitable for the medical domain, where interpretability is a main concern. The medical domain is data-intensive, and using electronic health record data to build the FRBS knowledge base and fuzzy sets is critical. Multiple variables are frequently required to determine a correct and personalized diagnosis, which usually makes it difficult to arrive at accurate and timely decisions. In this paper, we propose and implement a new semantically interpretable FRBS framework for diabetes diagnosis. The framework uses multiple aspects of knowledge-fuzzy inference, ontology reasoning, and a fuzzy analytical hierarchy process (FAHP) to provide a more intuitive and accurate design. First, we build a two-layered hierarchical and interpretable FRBS; then, we improve this by integrating an ontology reasoning process based on SNOMED CT standard ontology. We incorporate FAHP to determine the relative medical importance of each sub-FRBS. The proposed system offers numerous unique and critical improvements regarding the implementation of an accurate, dynamic, semantically intelligent, and interpretable CDSS. The designed system considers the ontology semantic similarity of diabetes complications and symptoms concepts in the fuzzy rules' evaluation process. The framework was tested using a real data set, and the results indicate how the proposed system helps physicians and patients to accurately diagnose diabetes mellitus This work was supported by National Research Foundation of Korea-Grant funded by the Korean Government (Ministry of Science, ICT and Future Planning)-NRF-2017R1A2B2012337) SI

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 Python 3.x module. The bridge between Python and Java is accomplished through the use of the Py4J framework. As a result, the possibility of using the IEEE standard for representing fuzzy systems is enlarged to a wider community of developers and knowledge engineers, with minimal code redundancy. Experiments show full interoperability between Python programs and JFML without any tangible overhead. We illustrate the use of Py4JFML in a beer style classification case study.

ExpliClas is a web service aimed at providing users with multimodal (textual + graphical) explanations related to Weka classifiers. In ExpliClas, two types of explanations are automatically generated. On the one hand, global explanations pay attention to the behavior of the classifier as a whole, i.e., they refer to a list of structural properties (number of classes, features, etc.) along with quality indicators such as accuracy or confusion matrix. On the other hand, local explanations go in depth with how the classifier deals with single instances. Current version of ExpliClas already explains classifications made by three different decision tree Weka implementations (J48, RepTree, and RandomTree) and one fuzzy algorithm (FURIA). In this paper, we describe ExpliClas in detail and illustrate its use with the Bank telemarketing dataset.

The existing studies of the interpretability of fuzzy systems have mainly focused on the analysis of the relation between a model of the fuzzy system and a human user considered as a beneficiary, i.e., a domain expert or a designer. Suggested by the concepts of “model” of formalized theories, “realization” of formalized language, and “interpretability” of a theory in another, the main contribution of this paper is the proposal of a new approach to the fuzzy system interpretability. This is grounded on the qualitative real-world semantics of words and relationships between the semantics of fuzzy system components and substructures of their real-world counterparts. Thus, we introduce a novel real-world-semantics-based approach. It is aimed at characterizing the so-called real-world-semantics-based interpretability of fuzzy systems. In addition, it considers the actual semantics of all fuzzy system components, including the inference engine. Moreover, this new approach opens the door to a new way to study the interpretability of fuzzy systems.

We present a novel heuristic approach that defines fuzzy geographical descriptors using data gathered from a survey with human subjects. The participants were asked to provide graphical interpretations of the descriptors ‘north’ and ‘south’ for the Galician region (Spain). Based on these interpretations, our approach builds fuzzy descriptors that are able to compute membership degrees for geographical locations. We evaluated our approach in terms of efficiency and precision. The fuzzy descriptors are meant to be used as the cornerstones of a geographical referring expression generation algorithm that is able to linguistically characterize geographical locations and regions. This work is also part of a general research effort that intends to establish a methodology which reunites the empirical studies traditionally practiced in data-to-text and the use of fuzzy sets to model imprecision and vagueness in words and expressions for text generation purposes., N/A

This paper presents an empirical research. It focuses on testing empirically the benefits of providing users, in a specific domain, with textual interpretation of the fuzzy inferences carried out by a fuzzy classifier for a given selection of samples. The hypothesis to test is as follows: “Users understand easier the decision made by a fuzzy system when they are provided with a textual interpretation of the fuzzy inference mechanism that the system carried out”. This hypothesis was successfully tested in a web survey. The application domain was leaf classification. The fuzzy classifiers were built with the GUAJE fuzzy modeling open source software which is aimed at generating interpretable fuzzy systems. The textual interpretation was handmade by an expert who followed the guidelines of the Natural Language Generation approach proposed by Reiter and Dale. Reported results encourage us to go on with a series of additional experiments devoted to deeply explore how Natural Language Generation techniques can contribute to facilitate the understanding of fuzzy systems.

The generation of text reports from numerical and symbolic data is getting the attention of many researchers. This paper presents an R package useful to develop computational systems able to generate linguistic descriptions of complex phenomena. It generates text reports from numerical and symbolic data related to the phenomena under consideration. This is an implementation of our previous research work that is supported by the computational theory of perceptions grounded in the fuzzy sets theory. Developing open source software while we follow the key issues (novelty, usability, interoperability, and relevance) will facilitate the adoption of this new discipline in the research community and industry. We present illustrative examples that show how to use this new R package. The examples reveal that the package is ready to become a relevant tool in the research field of text generation from data.

Translation is an important process for all living organisms. During translation, messenger RNA is rewritten into protein. Multiple control mechanisms determine how much protein is generated during translation. In particular, several regulatory elements located on mRNA transcripts are known to affect translation. In this study, a genome-wide analysis was performed to mine features related to translation in the genome of Arabidopsis thaliana. We used ribosome footprinting data to measure translation and constructed a predictive model using LASSO and random forest to select features that likely affect translation. We identified multiple transcript features and measured their influence on translation in different transcript regions. We found that features related to different translation stages may have a different impact on translation; often, features relevant to the elongation step were playing a stronger role. Interestingly, we found that the contribution of features may be different for transcripts belonging to different functional groups, suggesting that transcripts might employ different mechanisms for the regulation of translation.

Fuzzy systems have been used widely thanks to their ability to successfully solve a wide range of problems in different application fields. However, their replication and application require a high level of knowledge and experience. Furthermore, few researchers publish the software and/or source code associated with their proposals, which is a major obstacle to scientific progress in other disciplines and in industry. In recent years, most fuzzy system software has been developed in order to facilitate the use of fuzzy systems. Some software is commercially distributed, but most software is available as free and open-source software, reducing such obstacles and providing many advantages: quicker detection of errors, innovative applications, faster adoption of fuzzy systems, etc. In this paper, we present an overview of freely available and open-source fuzzy systems software in order to provide a well-established framework that helps researchers to find existing proposals easily and to develop well-founded future work. To accomplish this, we propose a two-level taxonomy, and we describe the main contributions related to each field. Moreover, we provide a snapshot of the status of the publications in this field according to the ISI Web of Knowledge. Finally, some considerations regarding recent trends and potential research directions are presented This work was supported in part by the Spanish Ministry of Economy and Competitiveness under Grants TIN2014-56633-C3-3-R and TIN2014-57251-P, the Andalusian Government under Grants P10-TIC-6858 and P11-TIC-7765, and the GENIL program of the CEI BioTIC GRANADA under Grant PYR-2014-2 SI

While some indoor Location Based Services (LBSs), such as medical equipment location in hospitals or people location in museums, do not need to estimate the trajectory of devices at short time intervals, some others, such as people guidance, require a frequent estimation of the device position. When providing an LBS for the latter, motion models and the information provided from motion sensors are commonly used to reduce the error in the localization, but this information is not always available. In this paper, we propose an approach to estimate the position of a moving device using a topological radio-map designed for static WiFi localization in a previous work. This approach uses a Bayes filter that continuously estimates the most likely position of the device. This filter will have to deal with the low working frequency of the device and the uncertainty of the observation to provide an accurate and fast estimation. Experiments performed in a real multi-floor environment show that the system is able to correctly track the device position, reducing the mean localization error.

This work presents the full integration of fuzzy inference-grams (Fingrams) in KEEL to visual analysis of fuzzy association rules. Fingrams graphically represent fuzzy rule-based systems (FRBSs) in 2D graphs that illustrate the interaction among fuzzy rules in terms of rule cofiring, i.e., paying attention to rule pairs simultaneous fired by a given input. The new module allows to generate Fingrams for fuzzy association rules created in the suite KEEL, that can be afterwards analyzed to comprehend the system behavior and improve it. We sketch the use and potentials in an illustrative example built in KEEL over a real-world dataset including qualitative assessments of a set of design chairs.

This paper presents an open source software tool for generation, visual representation and expert analysis of the so-called fuzzy inference-grams (fingrams in short). Fingrams are extended to fuzzy association rules in this work, expanding the scope of this methodology. The tool automatically generates fingrams and displays them to the user for expert analysis. It takes as input a simple configuration file that may be provided by whatever fuzzy modeling tool (it is actually provided by two widely known tools, GUAJE and KEEL). Fingrams graphically represent multi-dimensional fuzzy systems in 2D. The first version of this new software handles three different kinds of fuzzy systems (fuzzy rule-based systems for classification and regression, but also fuzzy association rules). Fingrams illustrate the interaction among fuzzy rules in terms of rule co-firing, i.e., paying attention to pairs of rules fired at the same time by a given input. Their visualization looks like a social network where nodes represent fuzzy rules while edges show the degree of interaction among pairs of rules. Thus, both representation and analysis are supported by social network analysis techniques. The utility of the new software tool is illustrated in a case study on visual representation and analysis of association rules.

Ribosome footprinting measures mRNA translation via deep sequencing of ribosome-protected mRNA fragments (footprints). Ribosome footprinting has been successfully applied to various model systems, including yeast, human, mouse, and worm; and it has been demonstrated that the technology is capable to produce global translation measurements with up to a single-codon resolution. Here we present our first experiences with a footprinting experiment that probes ethylene-induced translational changes in the reference plant Arabidopsis thaliana. The goal of our project is to develop genomic tools for the qualitative and quantitative characterization of the plant translatome. We hope ribosome footprinting will allow researchers to correlate precisely mRNA features with translation efficiency, and open the way to investigate globally the mechanism of protein synthesis and its regulation at an unprecedented level of detail.

This paper introduces a new system for people localization in indoor environments. It is mainly based on an intelligent classifier able to distinguish among a set of predefined and well identified topological locations. We adopt a WiFi signal strength fingerprint approach where most effort is done during the training phase, while online execution is fast and effective. Our proposal has been tested in a real environment with data collected in five different experimental sessions. Achieved results are encouraging since they overcome those ones provided by the well-known nearest neighbour fingerprint matching algorithm, that is usually considered as baseline for WiFi localization.

Although recently there has been many papers dealing with how to characterize and assess interpretability, there is still a lot of work to be done. Interpretability assessment is usually addressed by evaluating the complexity and/or readability of fuzzy rule-based systems. However, comprehensibility is usually not taken into account because it implies more cognitive aspects which are difficult to formalize and to deal with. In this work we show the importance of considering not only readability but also comprehensibility during the design process of fuzzy systems. We introduce the use of a novel index for evaluating comprehensibility in the context of a three-objective evolutionary framework for designing highly interpretable fuzzy rule-based classifiers. It is named as logical view index (LVI) and it is based on a semantic cointension approach. The proposed evolutionary algorithm consists of embedding the HILK (Highly Interpretable Linguistic Knowledge) fuzzy modeling methodology into the classical NSGA-II with the aim of maximizing accuracy, readability, and comprehensibility of the generated fuzzy rule-based classifiers. Our proposal is tested in the well-known PIMA benchmark problem which corresponds to a medical diagnosis problem where interpretability is a strong requirement.

Assessing interpretability of fuzzy systems still remains an open and challenging problem. Defining a good index is extremely difficult mainly due to the inherent subjective nature of interpretability. It strongly depends on the background of the person who makes the assessment according to its own knowledge, but also taking into account its previous experience and preferences. Since looking for fuzzy systems with a good accuracy-interpretability trade-off is required for many applications, guiding the whole design process by a good quality index would be extremely appreciated. Such index must be aware of both accuracy and interpretability. This paper introduces a framework that makes possible defining an index easily adaptable to the context of each problem by means of incorporating user's preferences and quality criteria. To do so, all aspects related to interpretability are first identified and then combined into a decision hierarchy framework. It is derived from a previous experimental study based on a web poll. The top of the hierarchy represents the quality index while the bottom includes all fuzzy systems to be evaluated. It consists of k decision levels structured as suggested by the classical analytic hierarchy process (AHP) defined by Saaty. In addition, the aggregation process is made using the ordered weighted averaging (OWA) operators defined by Yager. Such AHP+OWA combination was already proposed by Yager for solving multi-criteria decision problems. A simple example shows how the proposed method becomes effective but also efficient when assessing several fuzzy systems in an automatic process. The index is easily adaptable for providing those rankings expected by different users.

HILK (Highly Interpretable Linguistic Knowledge) is a fuzzy modeling methodology especially thought for designing interpretable fuzzy rule-based systems. As starting point, it trusts on a domain expert able to define the most influential variables along with the most suitable number of linguistic terms for each of them. However, such task is not easy because problems often involve too many variables. To tackle with this problem, present paper proposes embedding HILK in a three-objective evolutionary algorithm (HILKMO) with the aim of making genetic feature selection and fuzzy partition learning. The use of two-objective (maximizing accuracy and interpretability) evolutionary algorithm has become very popular and effective when dealing with modeling interpretable fuzzy systems. There are also works dealing with three objectives but two of them are usually related to interpretability regarding only the readability of the system description. We have already emphasized, in previous works, the importance of addressing also the system comprehensibility. Therefore, the main contribution of this work is introducing two contradictory goals for characterizing inter-pretability: maximizing readability of the system explanation. The former objective prefers rules as compact as possible, while the latter one favors the use of rules with low interaction among them because rule interaction is difficult to explain. Both objectives are contradictory because the more compact the rule base, the higher the chance of having rules simultaneously fired by the same input vector. We have chosen NSGA-II as multi-objective evolutionary algorithm and our proposal is tested in the well-known GLASS benchmark problem.

The goal of this paper is to study some of the most important WiFi signal variations, large and small scale variations and how they affect to WiFi localization systems. Moreover, the paper shows how to use Soft Computing techniques to deal with these uncertainties in WiFi localization systems. This work describes how to reduce uncertainty produced by small scale variations in indoor environments using fuzzy techniques. Some experimental results and conclusions are presented.

This work presents a new process for building comprehensible fuzzy systems for classification problems. Firstly, a feature selection procedure based on crisp decision trees is carried out. Secondly, strong fuzzy partitions are generated for all the selected inputs. Thirdly, a set of linguistic rules are defined combining the previously generated linguistic variables. Then, a linguistic simplification procedure guided by a novel interpretability index is applied to get a more compact and general set of rules without losing accuracy. Finally, an efficient and simple local search strategy increases the system accuracy while preserving the high interpretability. Results obtained in several benchmark classification problems are encouraging because they show the ability of the new methodology for generating highly interpretable fuzzy rule-based classifiers while yielding accuracy comparable to that achieved by other methods like neural networks and C4.5.

Traditionally, VF has been directly related to dispersion of refractoriness. However, further theoretical studies relating refractoriness and reentry generation would be helpful to understand their implication. Recently, the safety factor for conduction (SF) has been proposed as a useful parameter to analyze propagation of action potential (AP) and conduction block. In this work, the relative role of refractoriness and the source-sink relationship in conduction block leading to reentry during regional ischemia has been analyzed. Our results show that a) premature stimuli applied at CIs comprised between 170 ms and 204 ms lead to reentry, b) the distribution of SF during reentry is very similar for all the conducted simulations and c) in one third of the re-entries the premature stimulus was blocked at sites where the excitability of the cells was completely recovered.

Atrial fibrillation (AF) is the most common atrial tachyarrhythmia. Recently, pulmonary vein ectopic focus has been demonstrated that it can trigger reentry in the presence of a vulnerable substrate and lead to AF. We examined the effects of ectopic focus frequency on the generation of AF. In this study, the effects of remodelling on ionic currents were incorporated in a two-dimensional anisotropic model of human left atrial tissue. Ectopic activity initiated a stable reentry around the right pulmonary veins. The interaction of this ectopic activity with a rapid sinus rhythm generated fibrillatory conduction. This fibrillatory rhythm was maintained only at high rates of ectopic focus. Our study suggests that the interaction of high frequency ectopic activity and rapid sinus rhythm facilitates the progress from a stable reentrant mechanism to fibrillatory conduction in remodelling atrial tissue.

Grid computing technologies are mature enough to be successfully applied to computationally intensive scientific applications. However, the current process of applying grid computing to them is still hard and difficult for the least experienced users. In this paper we describe the adaptations made to the GMarte metascheduling framework in order to provide an ubiquitous access to its functionality as an efficient resource broker for the execution of tasks on computational grids. A metascheduling component accessible by only means of a Java-enabled Web browser has been developed, requiring almost zero configuration by the client. This approach has enabled to produce a generic multi-platform metascheduler which can be automatically deployed in the clients interested in resource brokering on computational grids based on the Globus Toolkit

With the recent advent of Grid Computing technologies, resource-starved applications can greatly benefit from the power that the Grid aims to deliver. In this paper, we focus on the application of Grid Computing to a biomedical computationally intensive application that simulates the electrical activity on cardiac tissues. A user level tool, called GMarte, was developed to simplify the usage of Grid facilities for the execution of scientific applications on distributed deployments. A cardiac case study has been executed, via GMarte, on the largest distributed Grid deployment available in Europe (the EGEE testbed) in order to assess the benefits of Grid Computing. The usage of a Grid infrastructure has dramatically reduced the time required to execute the case study, thus increasing research productivity.

Este articulo presenta una herramienta multiplataforma y multilingue, con una interfaz grafica amigable, que facilita la extraccion y representacion de conocimiento en sistemas basados en logica fuzzy. Su objetivo es la construccion o el refinamiento de bases de conocimiento integrando conocimiento experto e inducido, realizandose todo el proceso bajo el control del experto. La expresividad semantica de la logica fuzzy, proxima al lenguaje natural usado por el experto, simplifica el proceso de adquisicion y aumenta la interpretabilidad. / This paper presents a user-friendly portable tool designed and developed in order to make easier knowledge extraction and representation for fuzzy logic based systems. KBCT is an open source software that could be executed under Linux or Windows Operating Systems. Main goal of KBCT is the generation or refinement of fuzzy knowledge bases with a particular interest of obtaining interpretable partitions and rules. The use of fuzzy logic simplifies the knowledge extraction process and increase interpretability of rules because of the fuzzy rule expression is closed to expert natural language. KBCT lets the user define expert variables and rules, but also provide induction capabilities for partitions and rules. Both types of knowledge, expert and induced, are integrated under the expert control. In addition to this, the user can check consistency and quality of rule base at any moment. A simplify option is implemented in order to allow the user to reduce the size of rule base. The main objective consists of ensuring interpretability, non redundancy and consistency of the knowledge base along the whole process., Cette communication présente un logiciel multiplateforme et multilingue, avec une interface graphique développée, qui facilite l'extraction et la représentation de la connaissance sous la forme de systèmes basés sur la logique floue. Son objectif est la construction ou le raffinement de bases de connaissances qui intègrent de la connaissance experte comme de la connaissance induite. L'ensemble des procédures d'intégration est sous le contrôle de l'expert. L'utilisation de la logique floue, qui permet de modéliser le langage naturel, simplifie le processus d'acquisition et conduit à des systèmes hautement interprétables.

It is well known that short-long (S-L) sequences frequently precede the onset of malignant ventricular arrhythmias. Experimental studies have shown that a single long coupled extra-beat (L) after a previous basic cycle length resulted in a critical increase of the dispersion of repolarization pattern that promoted reentrant excitation. However, in normal intact heart little differences in repolarization among different myocardial layers have been observed after abrupt slowing of heart rate. We used the Luo and Rudy model (LRd00) of AP and the model of IKr blockade by the action of dofetilide previously developed by our group to study the effect of dofetilide (a class III antiarrhythmic drug) on the modulation of APD dispersion in a strand of ventricular tissue. Our results show that dofetilide increases dramatically the APD of the strand and that the effect of dofetilide on the modulation of dispersion is complex, indicating that the highest dispersion could appear before the highest effect of the drug is reached

The temperature dependence of the hysteretic and relaxational behaviours of a high coercivity La manganite was measured. The results were discussed in terms of their magnetic structure and they indicated that the vacancy-induced ferromagnetic Mn/sup 4+/-Mn/sup 3+/ clusters are coupled to the antiferromagnetic matrix in a way that involves decompensation of the antiferromagnetic lattice. It therefore resulted on the formation of net moments at the interfaces.