Object Membership Research Articles

Attribute reduction with fuzzy kernel-induced relations

Fuzzy rough sets have strong ability in dealing with data uncertainty, which have been widely applied in attribute reduction problems. Fuzzy relations are the cornerstones of fuzzy rough sets, and kernel-induced relations can simultaneously reflect the similarities of objects in both original space and high-dimensional Hilbert space. However, existing distance-based fuzzy relations (even kernel-induced) cannot truly reflect the correlations of objects under complex data distributions. To obtain better expressing ability for data, we propose a general optimization framework to get the memberships of objects with respect to categories. Considering various complex data distributions, the concepts of intra-class aggregation degree based on membership and inter-class dispersion degree based on nonmembership are proposed, and a separability measure is constructed to evaluate the importance of attributes to decisions. Correspondingly, a new attribute reduction algorithm is proposed, with a forward evaluation L-step strategy to skip local fluctuations in classification accuracy. Finally, we conduct a series of experiments to verify the effectiveness of the proposed algorithm. From the experiments on UCI, ELVIRA and face recognition datasets, it can be seen that the attribute subset obtained by the proposed algorithm not only has high classification accuracy and computational efficiency, but also has fewer attributes.

Find (and Remember) the Odd One Out: The Effect of Categorical Distinctiveness on Recognition Memory

How does the relationship between the categorical membership of a target object and its surrounding context affect the memorization of the target object? We conducted four experiments to examine how categorical distinctiveness is related to memory. During the categorical discrimination stage, participants searched for a target to be remembered among five distractor images belonging to another category. In the low categorical distinctiveness condition, the target image was surrounded by perceptually similar images from the same superordinate category (e.g., a cat among dogs); in the high categorical distinctiveness condition, it was surrounded by dissimilar objects from another superordinate category (e.g., a cat among chairs). Participants also performed a recognition test. We expected an increase in the number of hits and false alarms in the recognition task for objects discriminated in the low categorical distinctiveness condition. This hypothesis was confirmed in Experiments 1 and 4, and the effect also partly remained in Experiments 2 and 3 (when we manipulated the memorization time), affecting only hits or only false alarms, respectively. Moreover, in Experiment 2, in which memorization was incidental, memorization among low-distinguishability objects was more accurate, which could indicate the advantages that categorical knowledge offers in such a memorization paradigm. This indicates the influence of categorical knowledge on the memorization and recognition of objects.

Cross Entropy in Deep Learning of Classifiers Is Unnecessary-ISBE Error Is All You Need.

In deep learning of classifiers, the cost function usually takes the form of a combination of SoftMax and CrossEntropy functions. The SoftMax unit transforms the scores predicted by the model network into assessments of the degree (probabilities) of an object's membership to a given class. On the other hand, CrossEntropy measures the divergence of this prediction from the distribution of target scores. This work introduces the ISBE functionality, justifying the thesis about the redundancy of cross-entropy computation in deep learning of classifiers. Not only can we omit the calculation of entropy, but also, during back-propagation, there is no need to direct the error to the normalization unit for its backward transformation. Instead, the error is sent directly to the model's network. Using examples of perceptron and convolutional networks as classifiers of images from the MNIST collection, it is observed for ISBE that results are not degraded with SoftMax only but also with other activation functions such as Sigmoid, Tanh, or their hard variants HardSigmoid and HardTanh. Moreover, savings in the total number of operations were observed within the forward and backward stages. The article is addressed to all deep learning enthusiasts but primarily to programmers and students interested in the design of deep models. For example, it illustrates in code snippets possible ways to implement ISBE functionality but also formally proves that the SoftMax trick only applies to the class of dilated SoftMax functions with relocations.

Fuzzy-Rough Intrigued Harmonic Discrepancy Clustering

Fuzzy clustering decomposes data into clusters using partial memberships by exploring the cluster structure information, which demonstrates comparable performance for knowledge exploitation under the circumstance of information incompleteness. In general, this scheme considers the memberships of objects to cluster centroids and applies to clusters with the spherical distribution. In addition, the noises and outliers may significantly influence the clustering process; a common mitigation measure is the application of separate noise processing algorithms, but this usually introduces multiple parameters which are challenging to be determined for different data types. This paper proposes a new fuzzy-rough intrigued harmonic discrepancy clustering (HDC) algorithm by noting that fuzzy-rough sets offer a higher degree of uncertainty modelling for both vagueness and imprecision present in real-valued datasets. The HDC is implemented by introducing a novel concept of harmonic discrepancy, which effectively indicates the dissimilarity between a data instance and foreign clusters with their distributions fully considered. The proposed HDC is thus featured by a powerful processing ability on complex data distribution leading to enhanced clustering performance, particularly on noisy datasets, without the use of explicit noise handling parameters. The experimental results confirm the effectiveness of the proposed HDC, which generally outperforms the popular representative clustering algorithms on both synthetic and benchmark datasets, demonstrating the superiority of the proposed algorithm.

A Diffusion Model Analysis of Object-Based Selective Attention in the Eriksen Flanker Task.

Selective attention might be space-, feature-, and/or object-based. Clear support for the involvement of an object-based mechanism is rather scarce, possibly because the predictions of models from these different classes often overlap. Yet, only object-based models can account for a larger congruency effect (CE) in the Eriksen flanker task when flankers are more (vs. less) strongly grouped to the target, but spacing and other response-irrelevant features of target and flankers are held constant. Exactly this was observed by Kramer and Jacobson (1991). So far, this theoretically relevant finding has not been replicated closely. We replicated the finding in two web-based experiments. Specifically, CEs were larger when flanker lines were connected to the central target line (vs. to outer neutral lines). We also successfully fitted the Diffusion Model for Conflict tasks (DMC) to the experimental data. Critically, diffusion modeling (DMC) and distributional analyses (delta functions) revealed that object membership primarily affected target processing strength rather than strength or timing of flanker processing. This challenges the prominent attentional spreading (sensory enhancement) account of object-based selective attention and motivates an alternative target attenuation account.

A simple linear model to aid in analyses of the β Pictoris moving group

ABSTRACT We build a 4D linear model of object membership in the $\beta$ Pictoris moving group (BPMG), using two nested applications of Principal Component Analysis (PCA) for ∼1.5 million objects with high-quality data. These data consist of 3D-Galactic space velocities and Gaia G magnitudes. Through PCA, they ultimately result in a 4D straight line, referred to as PC 1′, about which both the bona fide members used to obtain the straight line and the candidate members used to test the model congregate at generally small distances. As a proof of concept, we select bona fide members from a recent, Gaia DR2-based compilation and candidate members mostly from a (pre-Gaia DR2) compilation from 2017. Using a standard procedure to flag groups of outliers in data sets, we find possible outliers flagged on account of their large distances to PC 1′, and find evidence that discrepant radial velocity measurements may indeed bring into question their validity as BPMG members. We propose that PC 1′ be added to the tool set for BPMG analyses and potentially extended to other young stellar moving groups.

The Massive and Distant Clusters of WISE Survey. XI. Stellar Mass Fractions and Luminosity Functions of MaDCoWS Clusters at z ∼ 1

We present stellar mass fractions and composite luminosity functions (LFs) for a sample of 12 clusters from the Massive and Distant Clusters of WISE Survey (MaDCoWS) in the redshift range 0.951 ≤ z ≤ 1.43. Using spectral energy distribution fitting of optical and deep mid-infrared photometry, we establish the membership of objects along the lines of sight to these clusters and calculate the stellar masses of member galaxies. This allows us to calculate the stellar mass of the clusters much more precisely than in previous works. We find stellar mass fractions for these clusters largely consistent with previous works, and an apparent negative correlation with total cluster mass. We measure a composite 3.6 μm LF down to m* + 2.5 for all 12 clusters. Fitting a Schechter function to the LF, we find a characteristic 3.6 μm magnitude of m* = 19.83 ± 0.12 and faint-end slope of α = −0.81 ± 0.10 for the full sample at a mean redshift of . We also divide the clusters into high- and low-redshift bins at and , respectively, and measure a composite LF for each bin. We see a small, but statistically significant, evolution in m* and α—consistent with passive evolution—when we study the joint fit to the two parameters, which is probing the evolution of faint cluster galaxies at z ∼ 1. This highlights the importance of deep IR data in studying the evolution of cluster galaxy populations at high redshift.

Hidden power of near-infrared data for the study of young clusters: Illustrative case of RCW 38

Studies of star formation rely heavily on observations in the near-infrared, but they typically need information from other wavelengths for interpretation. We show that we can infer distances and estimate the membership of young stellar objects for young clusters independently using (ground-based) near-infrared,J,H, andKSbroadband data alone. We also show that we can estimate a lower limit for the fraction of sources with 2.2 μm excess emission with a sensitivity comparable to that of mid-infrared space data, but with better resolution and fewer biases. Finally, we show that the typical methods for inferring masses from these data may produce substantially unreliable results. This method is applied to the young, massive cluster RCW 38, for which we estimate a distance of 1.5 kpc and aKS-band excess fraction larger than 60%.

When is ostensive communication used for joint action?

Abstract Joint actions typically require that information relevant for performing a task together is available to the interaction partners. In some situations, such information is perceptually retrievable and salient enough for co-actors to simply use it. In other situations, the relevant information needs to be actively shared among co-actors, e.g., by making it more perceptually salient or indicating it by means of a conventional signal. Here we consider a third case, where the information is not perceptually available and cannot be communicated by conventional means. How do joint action partners coordinate in such situations? We propose that co-actors resort to ostensive communication, that is, they draw attention to the fact that they intend to communicate some specific information. Two experiments tested the proposed role of ostensive communication for joint action. In a non-verbal joint building task, the category membership of different objects was known to only one person in a dyad, who needed to inform the partner which object type to use. In line with our hypothesis, most participants highlighted a particular object category with an ostensive gesture (characterized by containing more submovements than a natural placing movement) to resolve perceptual ambiguity. We conclude that ostensive communication is especially useful for joint action in situations where task-relevant information is not available to all co-actors and where it cannot be perceptually highlighted or conventionally communicated.

Remote Sensing Image Classification with a Graph-Based Pre-Trained Neighborhood Spatial Relationship.

Previous knowledge of the possible spatial relationships between land cover types is one factor that makes remote sensing image classification “smarter”. In recent years, knowledge graphs, which are based on a graph data structure, have been studied in the community of remote sensing for their ability to build extensible relationships between geographic entities. This paper implements a classification scheme considering the neighborhood relationship of land cover by extracting information from a graph. First, a graph representing the spatial relationships of land cover types was built based on an existing land cover map. Empirical probability distributions of the spatial relationships were then extracted using this graph. Second, an image was classified based on an object-based fuzzy classifier. Finally, the membership of objects and the attributes of their neighborhood objects were joined to decide the final classes. Two experiments were implemented. Overall accuracy of the two experiments increased by 5.2% and 0.6%, showing that this method has the ability to correct misclassified patches using the spatial relationship between geo-entities. However, two issues must be considered when applying spatial relationships to image classification. The first is the “siphonic effect” produced by neighborhood patches. Second, the use of global spatial relationships derived from a pre-trained graph loses local spatial relationship in-formation to some degree.

Decision-Making Approach with Fuzzy Type-2 Soft Graphs

Molodtsov’s theory of soft sets is free from the parameterizations insufficiency of fuzzy set theory. Type-2 soft set as an extension of a soft set has an essential mathematical structure to deal with parametrizations and their primary relationship. Fuzzy type-2 soft models play a key role to study the partial membership and uncertainty of objects along with underlying and primary set of parameters. In this research article, we introduce the concept of fuzzy type-2 soft set by integrating fuzzy set theory and type-2 soft set theory. We also introduce the notions of fuzzy type-2 soft graphs, regular fuzzy type-2 soft graphs, irregular fuzzy type-2 soft graphs, fuzzy type-2 soft trees, and fuzzy type-2 soft cycles. We construct some operations such as union, intersection, AND, and OR on fuzzy type-2 soft graphs and discuss these concepts with numerical examples. The fuzzy type-2 soft graph is an efficient model for dealing with uncertainty occurring in vertex-neighbors structure and is applicable in computational analysis, applied intelligence, and decision-making problems. We study the importance of fuzzy type-2 soft graphs in chemical digestion and national engineering services.

Fuzzy Divisive Hierarchical Clustering of Solvents According to Their Experimentally and Theoretically Predicted Descriptors

The present study describes a simple procedure to separate into patterns of similarity a large group of solvents, 259 in total, presented by 15 specific descriptors (experimentally found and theoretically predicted physicochemical parameters). Solvent data is usually characterized by its high variability, different molecular symmetry, and spatial orientation. Methods of chemometrics can usefully be used to extract and explore accurately the information contained in such data. In this order, advanced fuzzy divisive hierarchical-clustering methods were efficiently applied in the present study of a large group of solvents using specific descriptors. The fuzzy divisive hierarchical associative-clustering algorithm provides not only a fuzzy partition of the solvents investigated, but also a fuzzy partition of descriptors considered. In this way, it is possible to identify the most specific descriptors (in terms of higher, smallest, or intermediate values) to each fuzzy partition (group) of solvents. Additionally, the partitioning performed could be interpreted with respect to the molecular symmetry. The chemometric approach used for this goal is fuzzy c-means method being a semi-supervised clustering procedure. The advantage of such a clustering process is the opportunity to achieve separation of the solvents into similarity patterns with a certain degree of membership of each solvent to a certain pattern, as well as to consider possible membership of the same object (solvent) in another cluster. Partitioning based on a hybrid approach of the theoretical molecular descriptors and experimentally obtained ones permits a more straightforward separation into groups of similarity and acceptable interpretation. It was shown that an important link between objects’ groups of similarity and similarity groups of variables is achieved. Ten classes of solvents are interpreted depending on their specific descriptors, as one of the classes includes a single object and could be interpreted as an outlier. Setting the results of this research into broader perspective, it has been shown that the fuzzy clustering approach provides a useful tool for partitioning by the variables related to the main physicochemical properties of the solvents. It gets possible to offer a simple guide for solvents recognition based on theoretically calculated or experimentally found descriptors related to the physicochemical properties of the solvents.

A Heuristic-Based Model for MMMs-Induced Fuzzy Co-Clustering with Dual Exclusive Partition

MMMs-induced fuzzy co-clustering achieves dual partition of objects and items by estimating two different types of fuzzy memberships. Because memberships of objects and items are usually estimated under different constraints, the conventional models mainly targeted object clusters only, but item memberships were designed for representing intra-cluster typicalities of items, which are independently estimated in each cluster. In order to improve the interpretability of co-clusters, meaningful items should not belong to multiple clusters such that each co-cluster is characterized by different representative items. In previous studies, the item sharing penalty approach has been applied to the MMMs-induced model but the dual exclusive constraints approach has not yet. In this paper, a heuristic-based approach in FCM-type co-clustering is modified for adopting in MMMs-induced fuzzy co-clustering and its characteristics are demonstrated through several comparative experiments.

Prediction Based on the Solution of the Set of Classification Problems of Supervised Learning and Degrees of Membership

It is proposed to use the degrees of membership of objects to each class in the process of recognition in the linear corrector model to solve the problem of restoring dependences from precedent samples. Two models of the algorithm for calculating estimates are used as classifiers. The work of the proposed model is compared with the original method and with the well-known data analysis methods. The dependence of the work of the linear corrector on its parameters is studied.

Segmentation Assisted Object Distinction for Direct Volume Rendering

Ray Casting is a direct volume rendering technique for visualizing 3D arrays of sampled data. It has vital applications in medical and biological imaging. Nevertheless, it is inherently open to cluttered classification results. It suffers from overlapping transfer function values and lacks a sufficiently powerful voxel parsing mechanism for object distinction. In this work, we are proposing an image processing based approach towards enhancing ray casting technique for object distinction process. The rendering mode is modified to accommodate masking information generated by a K-means based hybrid segmentation algorithm. An effective set of image processing techniques are creatively employed in construction of a generic segmentation system capable of generating object membership information.

Bounded Fuzzy Possibilistic Method

This paper introduces the Bounded Fuzzy Possibilistic Method (BFPM) by addressing several issues that previous clustering/classification methods have not considered. In fuzzy clustering, the data object's membership values should sum to 1. Hence, any data object may obtain full membership in at most one cluster. Possibilistic clustering methods remove this restriction. However, BFPM differs from previous fuzzy and possibilistic clustering approaches by allowing the membership function to take larger values with respect to all clusters. Furthermore, in BFPM, a data object can have full membership in multiple clusters or even in all clusters. BFPM relaxes the boundary conditions (restrictions) in membership assignment. The assigned boundaries satisfy the necessity of obtaining full memberships and overcome the issues with conventional methods (overlapping). The functionality of the proposed method has been proved in geometry, set theory, and other disciplines, when learning methods should provide the important arithmetic operations on different domains. Weighted Feature Distance function (WFD) is also used to cover diversity. Validity and comparison indexes have been applied to evaluate the accuracy of BFPM. Conventional fuzzy and possibilistic methods are compared with BFPM and BFPM-WFD, in terms of accuracy, fuzzification constant, different norms, and covering diversity and overlapping. The results show that the proposed algorithms perform better than other fuzzy and possibilistic methods.

Liberalism and dictatorship in the problem of fuzzy classification

We investigate two related problems of categorization of alternatives. We are particularly concerned about liberalism and dictatorship, two noteworthy and opposing principles, in the context of fuzzy expressions of the opinions. First we study the problem of self-distributing the individuals in a society among fixed classes or categories. We identify a restriction on the fuzzy expression of the opinions that relies on Ruspini partitions, such that two characterizations of liberalism hold true. Then we show that similar axiomatizations can be produced with a more general structure of the fuzzy inputs, where the individuals are allowed to use ϵ-partitions. Secondly we study plausible sufficient conditions for the existence of dictators when a group of agents assigns degrees of membership of objects to a fixed list of classes.

Multi-objective reactive power optimization of hybrid AC/DC power system considering power system uncertainty

With the High Voltage Direct Current (HVDC) transmission is connected to the AC system, the uncertainty of converter reactive power consumption has brought certain influence to the reactive power optimization of the power grid. For the uncertainty of converter reactive power consumption and load, the DC reactive power and load are expressed by trapezoidal fuzzy parameters, and the multi-objective fuzzy reactive power optimization model of AC / DC interconnected system is established which optimizes the total network loss and node voltage offset of the AC / DC transmission system. The multi-objective optimization is converted into single-objective optimization by using the objective membership function and the fuzzy algorithm of multi-objective nonlinear objective programming, and the biogeography-based optimization algorithm is used to solve the model. The results of extended 39-node system example show that the proposed model and algorithm are suitable for voltage and reactive power control optimization under the condition that the converter reactive power consumption and load are fuzzy parameters.

Modified Recognition Algorithms Based on the Construction of Models of Elementary Transformations

This paper deals with the issues of building a model of recognition algorithms targeted at classifying objects in conditions of high dimensionality of the feature space. A model of recognition algorithms based on the estimate calculation has been considered as the source one. The distinctive features of the approach under consideration are as follows: 1) forming subsets of correlated features; 2) selecting a set of representative features when developing recognition algorithms; 3) building models of elementary transformations in the subspace of representative features. The main advantage of the proposed algorithms is the selection of the preferred models of elementary threshold rules with the subsequent calculation of the object membership estimate and ensuring a significant reduction in the number of computational operations when recognizing unknown objects. This characteristic is very important for real-time recognition systems. To test the operability of the proposed model, experimental studies have been carried out in solving a number of problems; in particular, model problems have been generated in the space of correlated features; as well as facial recognition tasks. These experimental studies considered: a model of recognition algorithms based on the estimate calculation; a model of recognition algorithms based on the potential function; the proposed model of recognition algorithms. A comparative analysis of the above models of recognition algorithms shows that the proposed model allows increasing the recognition accuracy by an average of 7-10%. The developed recognition algorithms can be used in medical and technical diagnostic, geological forecasting, biometric identification and other areas, in which solving problems of classification of objects specified in the high dimension feature space.

Prawne aspekty położenia mniejszości narodowych na Górnym Śląsku w latach 1918-1939

This paper discusses the legal situation of German and Polish ethnic minorities in Upper Silesia, which was divided between Germany and Poland in the wake of uprisings (1919-1921) and the plebiscite of March 1921. The discussion concentrates on the provisions of the Upper Silesia Convention (known also as the Geneva Convention) concluded by the German Reich and the Republic of Poland for fi fteen years under the auspices of the League of Nations in May 1922. Emphasis is laid on the main provisions, including the so-called fundamental rights of minorities (Art. 64-68), which were meant to ensure equal treatment and free development in the spheres of language, education, religion, etc. to ethnic minorities. The discussion also touches upon other issues – which were not fully regulated by the Convention – concerning the interpretation of the term ‘ethnic minority’ at the League of Nations and other organisations and institutions (Inter-Allied Mixed Commission for Upper Silesia), as well as in the prevailing legal opinions in Germany and Poland at that time. On the example of the views of such jurists as Bruns, Flachtbarth, Walz, Cybichowski, Kierski and Kostanecki, arguments and controversies are shown which surrounded the criteria for defi ning ethnic minorities. Over this matter two views clashed. The fi rst and more popular held a person to be member of an ethnic minority if he or she expressed their bona fi de will to be counted as one (subjective criterion). The second was based on the assumption of objective membership in an ethnic minority (criteria of language, religion, culture and tradition). In the author’s opinion, the Upper Silesia Convention contributed to the reduction of ethnic tensions in the area where it was enforced.