Binary Attributes Research Articles

Beta-CoRM: A Bayesian approach for n-gram profiles analysis

n-gram profiles have been successfully and widely used to analyse long sequences of potentially differing lengths for clustering or classification. Mainly, machine learning algorithms have been used for this purpose but, despite their predictive performance, these methods cannot discover hidden structures or provide a full probabilistic representation of the data. A novel class of Bayesian generative models designed for n-gram profiles used as binary attributes have been designed to address this. The flexibility of the proposed modelling allows to consider a straightforward approach to feature selection in the generative model. Furthermore, a slice sampling algorithm is derived for a fast inferential procedure, which is applied to synthetic and real data scenarios and shows that feature selection can improve classification accuracy.

Toward a Responsible Fairness Analysis: From Binary to Multiclass and Multigroup Assessment in Graph Neural Network-Based User Modeling Tasks

User modeling is a key topic in many applications, mainly social networks and information retrieval systems. To assess the effectiveness of a user modeling approach, its capability to classify personal characteristics (e.g., the gender, age, or consumption grade of the users) is evaluated. Due to the fact that some of the attributes to predict are multiclass (e.g., age usually encompasses multiple ranges), assessing fairness in user modeling becomes a challenge since most of the related metrics work with binary attributes. As a workaround, the original multiclass attributes are usually binarized to meet standard fairness metrics definitions where both the target class and sensitive attribute (such as gender or age) are binary. However, this alters the original conditions, and fairness is evaluated on classes that differ from those used in the classification. In this article, we extend the definitions of four existing fairness metrics (related to disparate impact and disparate mistreatment) from binary to multiclass scenarios, considering different settings where either the target class or the sensitive attribute includes more than two groups. Our work endeavors to bridge the gap between formal definitions and real use cases in bias detection. The results of the experiments, conducted on four real-world datasets by leveraging two state-of-the-art graph neural network-based models for user modeling, show that the proposed generalization of fairness metrics can lead to a more effective and fine-grained comprehension of disadvantaged sensitive groups and, in some cases, to a better analysis of machine learning models originally deemed to be fair. The source code and the preprocessed datasets are available at the following link: https://github.com/erasmopurif/toward-responsible-fairness-analysis.

Binary matrix factorization via collaborative neurodynamic optimization

Binary matrix factorization is an important tool for dimension reduction for high-dimensional datasets with binary attributes and has been successfully applied in numerous areas. This paper presents a collaborative neurodynamic optimization approach to binary matrix factorization based on the original combinatorial optimization problem formulation and quadratic unconstrained binary optimization problem reformulations. The proposed approach employs multiple discrete Hopfield networks operating concurrently in search of local optima. In addition, a particle swarm optimization rule is used to reinitialize neuronal states iteratively to escape from local minima toward better ones. Experimental results on eight benchmark datasets are elaborated to demonstrate the superior performance of the proposed approach against six baseline algorithms in terms of factorization error. Additionally, the viability of the proposed approach is demonstrated for pattern discovery on three datasets.

Are human categories formal concepts? A case study using Dutch data

While the notion of a formal concept, as used in formal concept analysis, is inspired by the traditional view of human concepts, the psychological relevance of formal concepts has not been examined in the past. In this paper, we provide an experimental exploration of the psychological plausibility of formal concepts as human categories. For this purpose, we use the currently most extensive available psychological data regarding human categories. The data involve several human categories, over 400 exemplars of these categories, several hundreds of binary attributes that describe these exemplars and several binary matrices representing which exemplars have which attributes. Our primary question is: Are human categories formal concepts? That is, do the involved human categories represent formal concepts in the respective exemplar-attribute binary matrices? In most of the examined instances, the answer to this question turns out affirmative. This supports the hypothesis that formal concepts provide a psychologically plausible model of human categories. In addition, we discuss several related questions, provide observations on the psychological data used and present topics for future exploration.

Exploration of polytomous-attribute Q-matrix validation in cognitive diagnostic assessment

Compared with typical binary attributes, polytomous attributes can take three or more values (corresponding to different levels of mastery of a respondent or measurement of an item). They can indicate whether a respondent possesses the attributes of interest and mastery levels. Therefore, the test with polytomous-attribute Q-matrix can become more informative and provide respondents with richer diagnostic information than the test based on the dichotomous-attribute Q-matrix. This paper extends the S-statistic and the residual method applicable for the Q-matrix of binary attributes to validate the polytomous-attribute Q-matrix. Under two common scenarios in real-world applications, two associated validation algorithms: the joint validation (JV) algorithm and the online validation (OV) algorithm, are proposed. Both simulation studies and an empirical data example were employed to assess the robustness and usefulness of these two methods under various conditions. Results indicate that the JV algorithm is suitable for validating a Q-matrix predefined by subject matter experts. Especially when the Q-matrix contains fewer misspecifications, while the OV algorithm can be applied to define the attribute vector of "new items". Based on a certain number of "operational items", the OV algorithm can achieve a promising performance for obtaining the specification of the new items.

Optimal choice designs for the main effects plus specified two-factor interaction effects model for binary attributes

Unlike optimal utility-neutral binary choice designs requiring a large number of choice sets to estimate the main effects and all two-factor interactions, we construct optimal binary choice designs with fewer choice sets for estimating the main effects plus specified two-factor interactions.

A general model for how attributes can reduce polarization in social groups

AbstractPolarization makes it difficult to form positive relationships across existing groups. Decreasing polarization may improve political discourse around the world. Polarization can be modeled on a social network as structural balance, where the network is composed of groups with positive links between all individuals in the group and negative links with all others. Previous work shows that incorporating attributes of individuals usually makes structural balance, and hence polarization, harder to achieve. That work examines only a limited number and types of attributes. We present a generalized model and a simulation framework to analyze the effect of any type of attribute, including analytically as long as an expected value can be written for the type of attribute. As attributes, we consider people’s (approximately) immutable characteristics (e.g., race, wealth) and such opinions that change more slowly than relationships (e.g., political preferences). We detail and analyze five classes of attributes, recapitulating the results of previous work in this framework and extending it. While it is easier to prevent than to destabilize polarization, we find that usually the most effective at both are continuous attributes, followed by ordered attributes and, finally, binary attributes. The effectiveness of unordered attributes varies depending on the magnitude of negative impact of having differing attributes but is smaller than of continuous ones. Testing the framework on network structures containing communities revealed that destroying polarization may require introducing local tensions. This model could be used by policymakers, among others, to prevent and design effective interventions to counteract polarization.

Fathoming the Mandela Effect: Deploying Reinforcement Learning to Untangle the Multiverse

Multiverse is a hypothetical idea that other universes can exist beyond our own. Various scientific theories have suggested scenarios such as the existence of bubble universes that constantly expand or string theory that attempts to merge gravity with other forces. Thus, a multiverse is a complex theoretical phenomenon that can best be conceived through computer simulation. Albeit within the multiverse, the causality of the Mandela effect is entirely possible. To examine the behavior of the multiverse as a representative ensemble, each universe as a specific ensemble element needs to be generated. Our universe generation is based on unique universes for two binary attributes of a population of n=303. The maximum possible universes this could produce within the multiverse is in the exponent of 182. To computationally confine the simulation to the scope of this study, the sample count of the multiverse is nmultiverse=606. Parameters representing the existence of each multiverse are implemented through the μ and σ values of each universe’s attributes. By using a developed reinforcement learning algorithm, we generate a multiverse yielding various universes. The computer gains consciousness of the parameters that can represent the expanse of possibility to exist for multiple universes. Furthermore, for each universe, a heart attack prediction model is performed to understand the universe’s environment and behavior. We test the Mandela effect or déjà vu of each universe by comparing error test losses with the training size of order M. Our model can measure the behavior of environments in different regions referred to as specific ensemble elements. By explicitly exploiting the attributes of each universe, we can get a better idea of the possible outcomes for the creation of other specific ensemble elements, as seen in the multiverse space planes.

Towards Zero-Shot Sign Language Recognition.

This paper tackles the problem of zero-shot sign language recognition (ZSSLR), where the goal is to leverage models learned over the seen sign classes to recognize the instances of unseen sign classes. In this context, readily available textual sign descriptions and attributes collected from sign language dictionaries are utilized as semantic class representations for knowledge transfer. For this novel problem setup, we introduce three benchmark datasets with their accompanying textual and attribute descriptions to analyze the problem in detail. Our proposed approach builds spatiotemporal models of body and hand regions. By leveraging the descriptive text and attribute embeddings along with these visual representations within a zero-shot learning framework, we show that textual and attribute based class definitions can provide effective knowledge for the recognition of previously unseen sign classes. We additionally introduce techniques to analyze the influence of binary attributes in correct and incorrect zero-shot predictions. We anticipate that the introduced approaches and the accompanying datasets will provide a basis for further exploration of zero-shot learning in sign language recognition.

Prediction Method of Human Group Emotion Perception Tendency Based on a Machine Learning Model

Humans with the same attribute tend to assume similar emotional and cognitive tendencies. In this study, a machine learning model was constructed to calculate the sample emotion categories’ membership matrices of human groups towards a group of aria samples, thus making a prediction of the human groups’ emotion perception tendencies. Subjects’ binary attributes in terms of gender, family environment and professional background were collected. The subjects were invited to perform an emotion identification of aria segments to build a multi-classification standard dataset of differences in subjects’ emotion identification. A dual-channel neural network model was proposed, and three neural networks based on it were adopted to predict the emotion perception tendencies of the categorical groups with three attributes. Subjects were clustered according to the predicted values of the emotion perception tendencies towards the sample test group, followed by a comparison of the clustering results with the actual attribute group classifications. As suggested by the experimental results, the neural network accurately predicted the classification groups’ emotion perception tendencies with the three attributes of the subjects. The cluster experiment results of subjects, based on the predicted values of the neural networks’ output, had better performance only under the professional background attribute.

On the Depth of Decision Trees with Hypotheses.

In this paper, based on the results of rough set theory, test theory, and exact learning, we investigate decision trees over infinite sets of binary attributes represented as infinite binary information systems. We define the notion of a problem over an information system and study three functions of the Shannon type, which characterize the dependence in the worst case of the minimum depth of a decision tree solving a problem on the number of attributes in the problem description. The considered three functions correspond to (i) decision trees using attributes, (ii) decision trees using hypotheses (an analog of equivalence queries from exact learning), and (iii) decision trees using both attributes and hypotheses. The first function has two possible types of behavior: logarithmic and linear (this result follows from more general results published by the author earlier). The second and the third functions have three possible types of behavior: constant, logarithmic, and linear (these results were published by the author earlier without proofs that are given in the present paper). Based on the obtained results, we divided the set of all infinite binary information systems into four complexity classes. In each class, the type of behavior for each of the considered three functions does not change.

Projecting social contact matrices to populations stratified by binary attributes with known homophily.

Contact networks are heterogeneous. People with similar characteristics are more likely to interact, a phenomenon called assortative mixing or homophily. Empirical age-stratified social contact matrices have been derived by extensive survey work. We lack however similar empirical studies that provide social contact matrices for a population stratified by attributes beyond age, such as gender, sexual orientation, or ethnicity. Accounting for heterogeneities with respect to these attributes can have a profound effect on model dynamics. Here, we introduce a new method, which uses linear algebra and non-linear optimization, to expand a given contact matrix to populations stratified by binary attributes with a known level of homophily. Using a standard epidemiological model, we highlight the effect homophily can have on model dynamics, and conclude by briefly describing more complicated extensions. The available Python source code enables any modeler to account for the presence of homophily with respect to binary attributes in contact patterns, ultimately yielding more accurate predictive models.

Inferring Latent Structure in Polytomous Data with a Higher-Order Diagnostic Model

Researchers continue to develop and advance models for diagnostic research in the social and behavioral sciences. These diagnostic models (DMs) provide researchers with a framework for providing a fine-grained classification of respondents into substantively meaningful latent classes as defined by a multivariate collection of binary attributes. A central concern for DMs is advancing exploratory methods for uncovering the latent structure, which corresponds with the relationship between unobserved binary attributes and observed polytomous items with two or more response options. Multivariate behavioral polytomous data are often collected within a higher-order design where general factors underlying first-order latent variables. This study advances existing exploratory DMs for polytomous data by proposing a new method for inferring the latent structure underlying polytomous response data using a higher-order model to describe dependence among the discrete latent attributes. We report a novel Bayesian formulation that uses variable selection techniques for inferring the latent structure along with a higher-order factor model for attributes. We report evidence of accurate parameter recovery in a Monte Carlo simulation study and present results from an application to the 2012 Programme for International Student Assessment (PISA) problem-solving vignettes to demonstrate the method.

Quota-based debiasing can decrease representation of the most under-represented groups

Many important decisions in societies such as school admissions, hiring or elections are based on the selection of top-ranking individuals from a larger pool of candidates. This process is often subject to biases, which typically manifest as an under-representation of certain groups among the selected or accepted individuals. The most common approach to this issue is debiasing, for example, via the introduction of quotas that ensure a proportional representation of groups with respect to a certain, often binary attribute. This, however, has the potential to induce changes in representation with respect to other attributes. For the case of two correlated binary attributes, we show that quota-based debiasing based on a single attribute can worsen the representation of the most under-represented intersectional groups and decrease the overall fairness of selection. Our results demonstrate the importance of including all relevant attributes in debiasing procedures and that more efforts need to be put into eliminating the root causes of inequalities as purely numerical solutions such as quota-based debiasing might lead to unintended consequences.

A heuristic scheme for multivariate set partitioning problems with application to classifying heterogeneous populations for multiple binary attributes

We provide a novel heuristic approach to solve a class of multivariate set partitioning problems in which each item is characterized by three attribute values. The scheme first identifies a series of orderings of the items and then solves a corresponding sequence of shortest path problems. We provide theoretical findings on the structure of an optimal solution that motivate the design of the proposed heuristic scheme. The proposed algorithm runs in polynomial-time and is independent of the number of groups in the partition, making it more efficient than existing algorithms. To measure the performance of our solutions, we construct bounds for special instances which allow us to provide optimality gaps. We conduct an extensive numerical experiment in which we solve a large number of problem instances and show that our proposed approach converges to the global optimal solution in the vast majority of cases and in the case it does not, it yields very low optimality gaps. We demonstrate our findings with an application in the context of classifying a large heterogeneous population as positive or negative for multiple binary attributes as efficiently as possible. We conduct a case study on the screening of three of the most prevalent sexually transmitted diseases in the United States. The resulting solutions are shown to be within 2.6% of optimality and lead to a 26% cost saving over current screening practices.

RoD: Evaluating the Risk of Data Disclosure Using Noise Estimation for Differential Privacy

Differential privacy is a paradigm of big data privacy protection that offers protection even when an attacker has arbitrary background knowledge in advance. Consequently, it is viewed as a reliable protection mechanism for sensitive information. Differential privacy introduces noise, such as Laplace noise, to obfuscate the true values in a data set while preserving its statistical properties. However, a large amount of Laplace noise added into a data set is typically defined by the discursive scale parameter of Laplace distribution. The privacy budget $\varepsilon$ ɛ in differential privacy has been theoretically interpreted, but the implication on the risk of data disclosure ( RoD ) in practice has not yet been well studied. Moreover, choosing an appropriate value for $\varepsilon$ ɛ is not straightforward because it considerably affects the level of privacy in a data set. In this paper, we define and evaluate RoD in a data set with either numerical or binary attributes for numerical or counting queries with multiple attributes based on noise estimation. Through confidence probability of noise estimation, we provide a simple method to select the privacy budget $\varepsilon$ ɛ for differential privacy and associate differential privacy with $k$ k -anonymization. Finally, we show the relationship between the RoD and $\varepsilon$ ɛ as well as between $\varepsilon$ ɛ and $k$ k in our experimental results. To the best of our knowledge, this is the first study using the quantity of noise as a bridge to evaluate RoD for multiple attributes (either numerical or binary data) and determine the relationship between differential privacy and $k$ k -anonymization.

MAAD-Face: A Massively Annotated Attribute Dataset for Face Images

Soft-biometrics play an important role in face biometrics and related fields since these might lead to biased performances, threaten the user's privacy, or are valuable for commercial aspects. Current face databases are specifically constructed for the development of face recognition applications. Consequently, these databases contain a large number of face images but lack in the number of attribute annotations and the overall annotation correctness. In this work, we propose a novel annotation-transfer pipeline that allows to accurately transfer attribute annotations from multiple source datasets to a target dataset. The transfer is based on a massive attribute classifier that can accurately state its prediction confidence. Using these prediction confidences, a high correctness of the transferred annotations is ensured. Applying this pipeline to the VGGFace2 database, we propose the MAAD-Face annotation database. It consists of 3.3M faces of over 9k individuals and provides 123.9M attribute annotations of 47 different binary attributes. Consequently, it provides 15 and 137 times more attribute annotations than CelebA and LFW. Our investigation on the annotation quality by three human evaluators demonstrated the superiority of the MAAD-Face annotations over existing databases. Additionally, we make use of the large number of high-quality annotations from MAAD-Face to study the viability of soft-biometrics for recognition, providing insights into which attributes support genuine and imposter decisions. The MAAD-Face annotations dataset is publicly available.

Localization of Mobile Robot by Particle Filter Considering Shape Information of Environment and Binary Attributes in Water Puddle

Localization of Mobile Robot by Particle Filter Considering Shape Information of Environment and Binary Attributes in Water Puddle

Advanced Nanosystems for Clinical Translation

Advanced Nanosystems for Clinical Translation

Attribute-Guided Multi-Scale Prototypical Network for Few-Shot SAR Target Classification

Few-shot synthetic aperture radar (SAR) target classification has received more and more attention in recent years, where most of the existing methods have applied existing networks designed for natural images to SAR images, ignoring the special characteristics of SAR data. Therefore, in this paper, we propose an attribute guided multi-scale prototypical network (AG-MsPN) combined with subband decomposition for few-shot SAR target classification, aiming to learn more discriminative features from a few labeled data. Since the SAR images are essentially complex-valued images containing both amplitude and phase information, we employ the subband decomposition on complex-valued SAR images to explore the backscattering behavior variation of targets to get more complete descriptions of targets. Then, considering the complementary information extracted by different convolutional layers, based on prototypical network, a multi-scale prototypical network is proposed to aggregate the features of different layers to enhance the discrimination of feature representations, thus relieving the problem of intraclass diversity and inter-class similarity. Besides, we devise the prior binary attributes of SAR targets and present the attribute classification module to further strengthen the performance of multi-scale prototypical network under the joint supervision of the label and prior attribute information. We demonstrate the effectiveness of our proposed AG-MsPN on the Moving and Stationary Target Acquisition and Recognition (MSTAR) benchmark dataset and our method surpasses many other existing methods in the few-shot cases.