Publications

Data science is a research field concerned with processes and systems that extract knowledge from massive amounts of data. In some situations, however, data shortage renders existing data-driven methods difficult or even impossible to apply. Transfer learning has recently emerged as a way of exploiting previously acquired knowledge to solve new yet similar problems much more quickly and effectively. In contrast to classical data-driven machine learning methods, transfer learning methods exploit the knowledge accumulated from data in auxiliary domains to facilitate predictive modeling in the current domain. A significant number of transfer learning methods that address classification tasks have been proposed, but studies on transfer learning in the case of regression problems are still scarce. This study focuses on using transfer learning techniques to handle regression problems in a domain that has insufficient training data. We propose an original fuzzy regression transfer learning method, based on fuzzy rules, to address the problem of estimating the value of the target for regression. A Takagi-Sugeno fuzzy regression model is developed to transfer knowledge from a source domain to a target domain. Experimental results using synthetic data and real-world datasets demonstrate that the proposed fuzzy regression transfer learning method significantly improves the performance of existing models when tackling regression problems in the target domain.

Cognitive Informatics (CI) and Cognitive Computing (CC) are fundamental intelligence theories and general AI technologies triggered by the transdisciplinary advances in intelligence, computer, brain, knowledge, cognitive, robotic, and cybernetic sciences for engineering implementations. This paper presents a summary of the plenary panel (Part I) on the theoretical foundations of CI/CC as well recent breakthroughs in AI engineering reported in the 20th IEEE International ICCI*CC Conference (ICCI*CC'21). The latest advances in CI and CC towards general AI are presented by twenty-two distinguished panelists. Strategic AI engineering applications in CI, CC, and cognitive systems are elaborated for abstract intelligence, general AI, cognitive robots, autonomous systems, intelligent vehicles, and safety-and-mission-critical systems.

An approach to adaptive user interface using mixture model and state space model is proposed. Mixture model is applied to response data of many users to extract user types in a preliminary experiment. Estimated components are regarded as "user types". Online identification of the type of a new user from his/her response series is done by state space model, where the weights of the components constitute the state vector. In the state space model, the system equation defines a time smoothness of the weights and the observation equation consists of a mixture model allocated to the time-varying weights. State estimation is done by using particle filter. We propose to use the identification result of the new user to an adaptive user interface by showing an appropriate screen based on the estimated weights. Numerical simulation illustrates type identification result of new user. Real data analysis using key-typing performance with methods using both-hands, right (dominant)-hand, left (non-dominant)-hand, and one finger is also reported.

In this study, we introduce a recent multicriteria decision theory concept of a new, generalized form of Choquet integral function and its application, in particular to the problem of face classification based on the aggregation of classifiers. Such function may be constructed by a simple replacement of the product used under the Choquet integral sign by any t-norm. This idea brings forward a broad class of aggregation operators, which can be incorporated into the decision-making theory. In this context, in a series of experiments we compare the most known t-norms and thoroughly examine their performance in the process of combining individual classifiers based either on facial regions or classic face recognition methods. Such kind of generalization can successfully improve the classification process provided that the parameters of the t-norms are carefully adjusted.

Granular Computing arose as a synthesis of insights into human-centred information processing by Zadeh in the late '90s and the Granular Computing name was coined, at this early stage, by T.Y Lin. Although the name is now in widespread use, or perhaps because of it, there are calls for a clarification of the distinctiveness of Granular Computing against the background of other human-centred information processing paradigms. This study examines the basic motivation for information granulation and casts Granular Computing as a structured combination of algorithmic and non-algorithmic information processing that mimics human, intelligent synthesis of knowledge from information.

Shadowed sets provide a meaningful description of information granules by abstracting the corresponding fuzzy sets into three categories: full acceptance, full rejection, and uncertain (represented by shadows). One of the main motivating points to derive shadowed sets from fuzzy sets is the determination and explanation of the separation thresholds based on a specific optimization mechanism. The available optimization objective functions are mainly discussed on semantic interpretations and their mathematical properties; constructive algorithms for optimal solutions have rarely been reported. In this paper, the continuous and convex properties of Pedrycz's optimization objective function to construct shadowed sets, as well as the existence and uniqueness of solution points, are analyzed in detail. It is demonstrated that different approximation region partitions would be generated even under the same optimization model, which requires further criteria to make the constructed shadowed sets well-defined. To address this limitation, the notions of passive and active constrained shadowed sets are introduced. A fast algorithm to obtain the proposed constrained shadowed sets is also designed based on the analyzed mathematical properties. Its performance is then illustrated by some typical fuzzy sets and some real data from the UCI repository.

In this paper, we develop a robust and fault-tolerant approach to microwave-based sensitive measurements using fuzzy neural network (FNN). Microwave chemic-identification, recently, is employing active planar ring resonators to enhance the resolutions significantly. However, in practice, when the technology of resolution improves, the results become more prone to minor variations in the measurement setup and user error. In order to eliminate these unwanted and uncontrollable deviations from the final allocations, we propose a novel and robust approach that uses more than one parameter out of measurements and incorporates FNN as a machine learning architecture at the post processing stage of sensing to obtain fault-tolerant classification. We have compared different membership functions used in the FNN and shown improvement in assigning accuracy from 49% (single parameter-dependent) up to 81.5% (three parameters-dependent) on an average of four materials, such as isopropanol-2 (IPA), ethanol, acetone, and water.

Multigranularity data analysis has recently become an active research topic in the intelligent computing and data mining fields. Feature selection via multigranularity data analysis is an effective tool for characterizing hierarchical data and enhancing the accuracy of the results. Although the multigranularity data analysis method has been widely adopted for feature selection, existing studies still present one prevalent disadvantage: multigranularity data analysis mostly focuses on information presented at a single granularity while ignoring the hierarchical structure of multigranularity data, which is contrary to the nature of multigranularity. Hence, this article proposes a multigranularity data analysis with a zentropy uncertainty measure for efficient and robust feature selection. Specifically, a consistent degree is first introduced to obtain optimal granularity combinations and establish an efficient neighborhood model for multigranularity information processing. Then, a novel and robust uncertainty measure is developed by integrating the multigranularity information, namely the zentropy-based measure. Considering its accuracy among uncertainty measures, two important measures are further designed and applied to feature selection. Extensive experiments demonstrate that the proposed method can achieve better robustness and classification performance than other state-of-the-art methods.

In this study, we discuss the use of fuzzy sets regarded as a well-rounded algorithmic vehicle in the construction of decision trees. The concept of fuzzy granulation realized via context-based clustering is aimed at the quantization (discretization) of continuous attributes as well as handling continuous classes encountered in classification problems. Two detailed experimental studies are presented concerning well-known data sets available on the Web.