This article is about explainable artificial intelligence (XAI) for rule-based fuzzy systems [that can be expressed generically, as$y({{\bf x}}) = f({{\bf x}})$]. It explains why it isnot validto explain the output of Mamdani or Takagi–Sugeno–Kang rule-based fuzzy systems using IF-THEN rules, and why itis validto explain the output of such rule-based fuzzy systems as anassociationof the compound antecedents of a small subset of the original larger set of rules, using a phrase such as “these linguistic antecedents aresymptomaticof this output.” Importantly, it provides a novel multi-step approach to obtain such a small subset of rules for three kinds of fuzzy systems, and illustrates it by means of a very comprehensive example. It also explains why the choice for antecedent membership function shapes may be more critical for XAI than before XAI, why linguistic approximation and similarity are essential for XAI, and, it provides a way to estimate the quality of the explanations.

iPresage is a web-based interactive tool for rapid and scalable patent landscaping. iPresage relies on sophisticated text mining, statistical machine learning and computational intelligence algorithms to allow analysts to mine large patent databases and evaluate whitespaces and temporal trends. This paper describes iPresage algorithms in detail and showcases iPresage functionality, using an illustrative example.

Imperfect information inevitably appears in real situations for a variety of reasons. Although efforts have been made to incorporate imperfect data into classification techniques, there are still many limitations as to the type of data, uncertainty and imprecision that can be handled. In this paper, we will present a Fuzzy Random Forest ensemble for classification and show its ability to handle imperfect data into the learning and the classification phases. Then, we will describe the types of imperfect data it supports. We will devise an augmented ensemble that can operate with others type of imperfect data: crisp, missing, probabilistic uncertainty and imprecise (fuzzy and crisp) values. Additionally, we will perform experiments with datasets used in other papers to show the advantage of being able to express the true nature of imperfect information.

Anomaly detection is a critical capability enabling condition-based maintenance (CBM) in complex real-world industrial systems. It involves monitoring changes to system state to detect “anomalous” behavior. Timely and reliable detection of anomalies that indicate faulty conditions can help in early fault diagnostics. This will allow for timely maintenance actions to be taken before the fault progresses and causes secondary damage to the system leading to downtime. When an anomaly is identified, it is important to isolate the source of the fault so that appropriate maintenance actions can be taken. In this paper, we introduce effective multivariate anomaly detection techniques and methods that allow fault isolation. We present experimental results from the application of these techniques to a high-bypass commercial aircraft engine.

When individual classifiers are combined appropriately, we usually obtain a better performance in terms of classification precision. Multi-classifiers are the result of combining several individual classifiers. In this work we propose and compare various combination methods to obtain the final decision of the multi-classifier based on a ldquoforestrdquo of randomly generated fuzzy decision trees, i.e., a Fuzzy Random Forest. We propose various forms of weighting decisions on the basis of information obtained from the FRF. We make a comparative study with several databases to show the efficiency of the various combination methods.

Fusing the outputs from an ensemble of models in an effective way can often boost overall model accuracy. This paper presents a novel method, called locally weighted fusion, which aggregates the results of multiple predictive models based on local accuracy measures of these models in the neighborhood of the probe point for which we want to make a prediction. While we demonstrate the method in the context of multiple neural network models, the concepts may be applied to other predictive techniques as well. This fusion method is applied to develop highly accurate models for emissions, efficiency, and load prediction in a complex real-world power plant. The locally weighted fusion method boosts the predictive performance by 20-40% over the baseline single model approach for the various prediction targets. Relative to this approach, fusion strategies which apply averaging or globally weighting only produce a 2-6% performance boost over the baseline.

Insurance underwriting is characterized as an ordinal classification problem since the underwriting process consists in assigning an application to one of an ordered set of risk categories. In designing ordinal classifiers, it is important to leverage the ordering information of the target classes to improve classification performance. In this paper, we explore several strategies for designing neural network based classifiers for ordinal classification. We investigate four different designs and evaluate their classification performance using real-world data from an automated insurance underwriting application.

We discuss implicit and explicit knowledge representation mechanisms for evolutionary algorithms (EAs). We also describe offline and online metaheuristics as examples of explicit methods to leverage this knowledge. We illustrate the benefits of this approach with four real-world applications. The first application is automated insurance underwriting-a discrete classification problem, which requires a careful tradeoff between the percentage of insurance applications handled by the classifier and its classification accuracy. The second application is flexible design and manufacturing-a combinatorial assignment problem, where we optimize design and manufacturing assignments with respect to time and cost of design and manufacturing for a given product. Both problems use metaheuristics as a way to encode domain knowledge. In the first application, the EA is used at the metalevel, while in the second application, the EA is the object-level problem solver. In both cases, the EAs use a single-valued fitness function that represents the required tradeoffs. The third application is a lamp spectrum optimization that is formulated as a multiobjective optimization problem. Using domain customized mutation operators, we obtain a well-sampled Pareto front showing all the nondominated solutions. The fourth application describes a scheduling problem for the maintenance tasks of a constellation of 25 low earth orbit satellites. The domain knowledge in this application is embedded in the design of a structured chromosome, a collection of time-value transformations to reflect static constraints, and a time-dependent penalty function to prevent schedule collisions.

A principal challenge in modern computational finance is efficient portfolio design - portfolio optimization followed by decision-making. Optimization based on even the widely used Markowitz two-objective mean-variance approach becomes computationally challenging for real-life portfolios. Practical portfolio design introduces further complexity as it requires the optimization of multiple return and risk measures subject to a variety of risk and regulatory constraints. Further, some of these measures may be nonlinear and nonconvex, presenting a daunting challenge to conventional optimization approaches. We introduce a powerful hybrid multiobjective optimization approach that combines evolutionary computation with linear programming to simultaneously maximize these return measures, minimize these risk measures, and identify the efficient frontier of portfolios that satisfy all constraints. We also present a novel interactive graphical decision-making method that allows the decision-maker to quickly down-select to a small subset of efficient portfolios. The approach has been tested on real-life portfolios with hundreds to thousands of assets, and is currently being used for investment decision-making in industry.

We analyze the task of selecting the most reliable units within a fleet of vehicles and formulate it as a prediction and classification problem. The prediction of each unit's remaining life is based on the identification of "peer" units, i.e. vehicles with similar utilization and maintenance records that are expected to behave similarly to the unit under consideration. With these peers, we construct local predictive models to estimate the unit's remaining life. We use evolutionary algorithms (EAs) to develop the criteria for defining peers and the relevance of each criterion in evaluating similarity with the unit. Each individual in the EAs population fully characterizes an instance-based fuzzy model that is used to predict the unit's remaining life. The precision of the selection of units with best-expected life provides the fitness value. We analyzed the performance of the evolutionary approach over two years of operation and maintenance data for a fleet of 1100 locomotives. The results illustrate the high accuracy and robustness of this approach. In the conclusion, we highlight the implications of this approach for supporting the lifecycle of the fuzzy models

In recent years, multi-objective evolutionary algorithms (MOEA) have generated a large research interest. MOEA's attraction stems from their ability to find a set of Pareto solutions rather than any single, aggregated optimal solution for a multi-objective problem. As for single-objective evolutionary algorithms (SOEA), multi-objective evolutionary algorithms also require parameter tuning to achieve desirable performance. In the literature we can find fuzzy logic controllers (FLC's) applied to online parameter control for SOEA. In this paper, we propose to use a FLC to dynamically adjust the parameters of a particular multi-objective differential evolution (MODE) algorithm. The fuzzy logic controlled multi-objective differential evolution (FLC-MODE) is applied to a suite of benchmark functions. Its results are compared to those obtained by using MODE with constant parameter settings. We show that the FLC-MODE obtains better results in 80% of the testing examples. Given that the benchmarks were synthetic test functions, we designed the FLC using only our understanding of the working mechanism of the MODE, without incorporating any additional problem-specific knowledge. When addressing real-world applications, we expect the FLC to be an excellent way for representing and leveraging their associated heuristic knowledge

This paper explores a novel ensemble technique for unsupervised classification using nonparametric statistics. Multiple classification systems (MCS), or ensemble techniques, involve considering several classification methods or multiple outputs from the same method and devising techniques to reach a decision. The performance of a binary classification system can be measured on a receiver operating characteristic (ROC) curve, and the area under the curve (AUC) is exactly the Wilcoxon rank sum or Mann-Whitney U statistic, both of which are nonparametric statistics based upon ranked data. Successful performance of an unsupervised ensemble can be measured through the AUC, and the performance of different aggregation techniques for the combination of the multiple classification system decision values, or rankings in this paper, is illustrated. Aggregation techniques are based upon fuzzy logic theory, creating the fuzzy ROC curve. The one-class SVM is utilized for the unsupervised classification.

The authors are of the opinion that building a good application calls, first and foremost, a profound understating of the application domain with all of its (sometimes highly complex) subtleties in order that a comprehensive knowledgebase can be embedded in the tool because that is what will provide a good answer and not the refinements of the underlying mathematical system.

A retrospective argument supporting the adaptive control of the resources of an evolutionary algorithm is presented. In this approach, a fuzzy controller monitors the population diversity and maturity of the evolutionary process, and issues control actions to adapt the algorithm's population size and mutation rate. The fuzzy controller's rule base is simple and intuitive, and encodes expert heuristic knowledge on the management of an evolutionary algorithm's resources. Analysis based on experimentation and statistical hypothesis testing reveals that such an adaptive approach significantly reduces the variance in search performance of the evolutionary algorithm. Experiments also reveal that the approach in general improves search performance of the algorithm.

A robust method for automating the tuning and maintenance of fuzzy decision-making systems is presented. A configurable multi-stage mutation-based evolutionary algorithm optimally tunes the decision thresholds and internal parameters of fuzzy rule-based and case-based systems that decide the risk categories of insurance applications. The tunable parameters have a critical impact on the coverage and accuracy of. decision-making, and a reliable method to optimally tune these parameters is critical to the quality of decision-making and maintainability of these systems.

Many automobile accidents could be avoided if the correct amount of brake pressure had been applied at the proper time. In this paper we propose the use of a fuzzy control system that could be embedded in an automobile to remove the driver's reaction time from the situation. The system, consisting of two fuzzy controllers connected in cascade, analyzes possible accident situations based on the vehicle separation distance, their relative velocity and a static friction coefficient that represents the road conditions. The controller determines the sufficient brake pressure required to prevent collisions while providing a smooth ride for the vehicle's passengers. This smooth performance was achieved by using genetic algorithms to tune the fuzzy controllers' parameters.

A methodology is described for an automated decision making process. The output of the decision process is a solution to the decision and a confidence value in the correctness of the solution. A fuzzy case-based reasoning approach is used to determine the solution. A fuzzy confidence engine uses values that are by-products of determining the solution in order to calculate the confidence value. This methodology has been implemented in a number of applications.

Soft computing (SC) is an association of computing methodologies that includes as its principal members fuzzy logic, neurocomputing, evolutionary computing and probabilistic computing. We present a collection of methods and tools that can be used to perform diagnostics, estimation, and control. These tools are a great match for real-world applications that are characterized by imprecise, uncertain data and incomplete domain knowledge. We outline the advantages of applying SC techniques and in particular the synergy derived from the use of hybrid SC systems. We illustrate some combinations of hybrid SC systems, such as fuzzy logic controllers (FLCs) tuned by neural networks (NNs) and evolutionary computing (EC), NNs tuned by EC or FLCs, and EC controlled by FLCs. We discuss three successful real-world examples of SC applications to industrial equipment diagnostics, freight train control, and residential property valuation.

Fuzzy logic control (FLC) technology has drastically reduced the development time and deployment cost for the synthesis of nonlinear controllers for dynamic systems. As a result we have experienced an increased number of FLC applications. We illustrate some of our efforts in FLC technology transfer, covering projects in turboshaft aircraft engine control, steam turbine startup, steam turbine cycling optimization, resonant converter power supply control, and data-induced modeling of the nonlinear relationship between process variables in a rolling mill stand. We compare these applications in a cost/complexity framework, and examine the driving factors that led to the use of FLCs in each application. We emphasize the role of fuzzy logic in developing supervisory controllers and in maintaining explicit tradeoff criteria used to manage multiple control strategies. Finally, we describe some of our FLC technology research efforts in automatic rule base tuning and generation, leading to a suite of programs for reinforcement learning, supervised learning, genetic algorithms, steepest descent algorithms, and rule clustering.>

This paper presents a new approach to planning in dynamic and uncertain environments. Planning is viewed as a process in which an agent's long term goals are transformed into short term tasks and objectives, given the agent's strategy and the context of planning. The developed model allows for a dynamic balance between long term strategic planning and short term tactical planning. The notion of a strategy hierarchy is introduced to explicitly represent the process of strategy formulation and refinement. A combination of rules and plan scripts is used. Rules are used for representing domain knowledge and reasoning about strategic choices. Plan scripts are used for representing specific tasks and objectives. The uncertainty calculi of RUM/PRIMO are used for supporting reasoning under uncertainty. In the proposed model, it is also possible to achieve a seamless integration of case-based reasoning into the planning process. These ideas have been implemented in a system called MARS, which plans in the financial domain of mergers and acquisitions.>

The representation of cases and rules is addressed. A case representation language tool used to directly represent instantiated cases is presented. PRIMO (plausible reasoning module), a rule-based system used to represent uncertainty and vagueness in data and rules, is described. Cases and rules, integrated in a combined approximate reasoning system (CARS), are illustrated with examples taken from the domain of mergers and acquisitions.>

A technique for making linguistic decsions is presented. Fuzzy sets are assued to be an appropriate way of dealng with uncertainty, and it b therefore cncluded that decisions taken on the basis of such infomation must themselves be fuzzy. It b inappriate then to present the decision in numencal form; a statement in natural angug is much better. For brevity only a single-stage multlabute decsion problem is considered. Solutions to such problems are shown using ideas in linguisc approximation and truth qualiction. An extensive example illuminates the basic ideas and techniques.