Bipartite Graph Model Research Articles

Personalized Activity Recognition Using Partially Available Target Data

Recent years have witnessed a growing body of research on autonomous activity recognition models for use in deployment of mobile systems in new settings such as when a wearable system is adopted by a new user. Current research, however, lacks comprehensive frameworks for transfer learning. Specifically, it lacks the ability to deal with partially available data in new settings. To address these limitations, we propose <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">OptiMapper</i> , a novel uninformed cross-subject transfer learning framework for activity recognition. OptiMapper is a combinatorial optimization framework that extracts abstract knowledge across subjects and utilizes this knowledge for developing a personalized and accurate activity recognition model in new subjects. To this end, a novel community-detection-based clustering of unlabeled data is proposed that uses the target user data to construct a network of unannotated sensor observations. The clusters of these target observations are then mapped onto the source clusters using a complete bipartite graph model. In the next step, the mapped labels are conditionally fused with the prediction of a base learner to create a personalized and labeled training dataset for the target user. We present two instantiations of OptiMapper. The first instantiation, which is applicable for transfer learning across domains with identical activity labels, performs a one-to-one bipartite mapping between clusters of the source and target users. The second instantiation performs optimal many-to-one mapping between the source clusters and those of the target. The many-to-one mapping allows us to find an optimal mapping even when the target dataset does not contain sufficient instances of all activity classes. We show that this type of cross-domain mapping can be formulated as a transportation problem and solved optimally. We evaluate our transfer learning techniques on several activity recognition datasets. Our results show that the proposed community detection approach can achieve, on average, 69 percent utilization of the datasets for clustering with an overall clustering accuracy of 87.5 percent. Our results also suggest that the proposed transfer learning algorithms can achieve up to 22.5 percent improvement in the activity recognition accuracy, compared to the state-of-the-art techniques. The experimental results also demonstrate high and sustained performance even in presence of partial data.

Joint Optimization for Coordinated Charging Control of Commercial Electric Vehicles Under Distributed Hydrogen Energy Supply

The transition to the zero-carbon power system is underway accelerating recently. Hydrogen energy and electric vehicles (EVs) are promising solutions on the supply and demand sides. This brief presents a novel architecture that includes hydrogen production stations, fast-charging stations, and commercial EVs. The proposed architecture jointly optimizes the hydrogen energy dispatch and the EV charging location selection and is formulated by a time-varying bilevel bipartite graph (T-BBG) model. We develop a bilevel iteration optimization method combining the linear programming (LP) and the Kuhn–Munkres (KM) algorithm to solve the joint problem whose optimality is proved theoretically. The effectiveness of the proposed architecture on reducing the operating cost is verified via case studies in Shanghai. The proposed method outperforms other strategies and improves the performance by at least 13%, which shows the potential economic benefits of the joint architecture. The convergence and impact of parameters are assessed.

Modeling Multi-Year Customers’ Considerations and Choices in China’s Auto Market Using Two-Stage Bipartite Network Analysis

Choice modeling is important in transportation planning, marketing and engineering design, as it can quantify the influence of product attributes and customer demographics on customers’ choice behaviors. Consumer studies suggest that customers’ choice-making process often consists of two different stages: customers first consider subsets of available products on the market, and then make the final choice from the subsets. As existing preference modeling is mostly focused on the choice stage, there is a need to develop methods for understanding customer preferences at both stages, and investigate how customer preferences change from “consideration” to “choice”, and whether such changes will be consistent over time. In this paper, we study customers’ consideration and purchase behaviors in China’s auto market using multi-year survey datasets. We demonstrate how descriptive network analysis and analytic network models (bipartite Exponential Random Graph Model (ERGM)) capture the change of customers’ preferences from the consideration stage to the choice stage in multiple consecutive years. Our results show that factors such as fuel consumption per unit power, car make origin, and place of production influence customers’ considerations and final purchase decisions in different ways, and this difference between consideration and purchase is consistent over time. The main contribution of this study is that we validate the two-stage network-based modeling approach and its utility in preference elicitation using multiple-year dataset, which sheds lights on understanding the trend of customers’ consideration and choice behaviors across years. Our study also contributes to a refined interpretation of the ERGM results with categorization of continuous variables into ranges, which shows that customer choice decisions may be more qualitatively influenced by product attributes rather than quantitatively. Our approach is generic and thus can be applied to solving broader choice modeling problems, such as the transportation mode selection and the adoption of clean technology (e.g., electric vehicles).

Donor interactions in the allocation of adaptation aid: A network analysis

This article examines how adaptation aid is allocated across countries, and specifically focus on the role of donor—donor interactions in allocation decisions. We test two contrasting hypotheses: the presence of other adaptation donors in a recipient country may increase or reduce the likelihood of donor i to provide adaptation aid to that recipient. In the former case, donors support adaptation in the same recipient countries; in the latter, they provide their adaptation aid to different recipient countries. We model adaptation aid allocations as a network, and apply an innovative method, bipartite temporal exponential random graph models, to bilateral adaptation aid flows between 2010 and 2016. Our empirical analysis finds strong evidence for donor interactions. The results suggest a positive effect of other donors: donors tend to support adaptation in similar sets of recipient countries. These results provide further evidence that adaptation aid largely follow the structures and processes of traditional development aid, which poses questions for the additionality of finance for adaptation to climate change.

High-throughput secure multiparty multiplication protocol via bipartite graph partitioning

For the privacy-preserving computation of multi-source large scale data sets, the secure multi-party computation protocol with high-throughput is of the utmost importance. However, the existing high-throughput secure multi-party protocols only involve the fixed 3-party or 4-party setting, limiting its practicality. To achieve a high-throughput n-party (n ≥ 3) secure protocol, low communication and simple computation are two major issues to be considered, which can be used to reduce network load and increase concurrency processing. In this paper, we design a secure multi-party multiplication protocol with only a single round interaction and simple computation by using replicated sharing, which is generated according to the partition of all cross-terms in the sharing-based multiplication operation. Furthermore, in order to implement the optimal communication for each round, we model all cross-terms of the sharing-based multiplication operation as a bipartite graph, and propose a bipartite graph partitioning algorithm. Due to the bipartite graph model, the optimal partition of the cross-terms can be reduced to partition the bipartite graph into n independent subgraphs with the least number of vertices in each subgraph. Finally, the evaluation results show the proposed protocol is both low communication and simple computation. In the case of the 4-party setting Boolean circuits, it only needs to send 1.5 bits and carry out 4 AND and 3 XOR operations on average per AND gate for each party, and achieving a rate of over 0.65 million AES per second.

Clustering spatio-temporal bi-partite graphs for finding crowdsourcing communities in IoMT networks

ABSTRACT The Internet of Moving Things is rapidly becoming a reality where intelligent devices and infrastructures are fostering real-time data sustainability in smart cities and advancing crowdsourced tasks to improve energy consumption, waste management, and traffic operations. These intelligent devices create a complex network scenario in which they often move together or in conjunction with one another to complete crowdsourced tasks. Our research premise is that mobility relationships matter when performing these tasks, and therefore, a graph model based on representing the changes in mobility relationships is needed to help identify the neighbour devices that are moving close to one another in our physical world but also seamlessly connected in their virtual world. We propose a bi-partite community mobility graph model for linking intelligent devices in both virtual and physical worlds, as well as reaching a trade-off between crowdsourced tasks designed with explicit and implicit citizen participation. This paper aims to explore a bi-partite graph as a promising spatio-temporal representation of IoMT networks since changes in mobility relationships over time can indicate volunteer organisation at the device and community levels. The Louvain community detection method is proposed to find communities of intelligent devices to reveal a value conscious participation of citizens. The proposed bi-partite graph model is evaluated using a real-world scenario in transportation, confirming the main role of evolving communities in developing crowdsourcing IoMT networks.

A Course Teacher Recommendation Algorithm Based on Improved Latent Factor Model and PersonalRank

To scientifically and accurately recommend suitable teachers for university courses and improve teaching quality, designing an effective recommendation algorithm is necessary. Therefore, we construct quantitative models of teacher characteristics, course characteristics, and teaching evaluations under the theories and methods of education and build a sparse experimental data matrix based on the quantified data. On this basis, we propose a teacher recommendation algorithm (PRLFM) based on the improved latent factor model (LFM) and the improved PersonalRank algorithm. Firstly, the improved LFM is used to predict the evaluation scores of those courses that teachers have not taught. The scores which are higher than the specified threshold are used to fill the corresponding missing items in the sparse matrix to reduce the matrix's sparsity. Then, the bipartite graph model based on the teacher set and course set is constructed according to the filled experimental data matrix. The weight of edges in the bipartite graph is replaced by the teacher and course's evaluation score multiplied by the course difficulty, which reflects the correlation between course and evaluation score. Next, an improved probability transition matrix based on the bipartite graph is constructed. The access probability in the matrix is replaced by the node's out degree's reciprocal multiplied by the edge's weight. The correlation degree between the course and all teachers is quickly calculated using the matrix algorithm of PersonalRank. Finally, a teacher recommendation model is constructed to realize teachers' top-N recommendation by combining the correlation degree with teachers' characteristics. Experiments show that the PRLFM algorithm can effectively improve the accuracy of prediction and top-N recommendation. It solves the problem of lack of scientific basis in recommending suitable teachers for university courses and improving the teaching quality.

An Automated Penetration Semantic Knowledge Mining Algorithm Based on Bayesian Inference

Mining penetration testing semantic knowledge hidden in vast amounts of raw penetration testing data is of vital importance for automated penetration testing. Associative rule mining, a data mining technique, has been studied and explored for a long time. However, few studies have focused on knowledge discovery in the penetration testing area. The experimental result reveals that the long-tail distribution of penetration testing data nullifies the effectiveness of associative rule mining algorithms that are based on frequent pattern. To address this problem, a Bayesian inference based penetration semantic knowledge mining algorithm is proposed. First, a directed bipartite graph model, a kind of Bayesian network, is constructed to formalize penetration testing data. Then, we adopt the maximum likelihood estimate method to optimize the model parameters and decompose a large Bayesian network into smaller networks based on conditional independence of variables for improved solution efficiency. Finally, irrelevant variable elimination is adopted to extract penetration semantic knowledge from the conditional probability distribution of the model. The experimental results show that the proposed method can discover penetration semantic knowledge from raw penetration testing data effectively and efficiently.

Rainbow k-connectivity of Random Bipartite Graphs

A path in an edge-colored graph G is called a rainbow path if no two edges of the path are colored the same color. The minimum number of colors required to color the edges of G such that every pair of vertices are connected by at least k internally vertex-disjoint rainbow paths is called the rainbow k-connectivity of the graph G, denoted by rck(G). For the random graph G(n,p), He and Liang got a sharp threshold function for the property rck(G(n,p)) ≤ d. For the random equi-bipartite graph G(n,n,p), Fujita et. al. got a sharp threshold function for the property rck(G(n,n,p)) ≤ 3. They also posed the following problem: For d ≥ 2, determine a sharp threshold function for the property rck(G) ≤ d, where G is another random graph model. This paper is to give a solution to their problem in the general random bipartite graph model G(m,n,p).

Network Modeling of Assisted Living Facility Residents' Attendance at Programmed Group Activities: Proximity and Social Contextual Correlates of Attendance.

Social engagement, including participation in group activities, supports older adults' mental and physical health. However, many residents of assisted living facilities do not participate in their facility's programmed group activities. Explaining residents' attendance at group activities is complex; attendance is associated with a confluence of individual-level and contextual factors. The aim of this study was to assess the effects of multilevel factors on attendance, including residents' proximity to activity location and the potential for one resident's attendance to depend on other residents' attendance. We used bipartite exponential random graph models to examine the attendance of 35 residents at 563 programmed group activities. We simultaneously modeled the effects of the geospatial distance from a resident's apartment to the activity and the tendency for residents to attend activities with similar groups of other residents (i.e., shared attendance) on the likelihood of attendance, while controlling for individual-level factors (demographic and health indicators) and activity popularity. Greater distance was associated with a reduced likelihood of attendance (p < .001) and residents tended to attend activities with similar subsets of other residents (p < .001). Findings suggest that greater distance to group activities may be a barrier to attendance. Implications include facility design, placement of activities and apartments, and mobility-related strategies to increase activity participation. It is also important to recognize that participation in activities is dependent on social context. We recommend that proximity to activities and social contextual factors be considered in future examinations of attendance at group activities.

A dynamic programming approach for distributing quantum circuits by bipartite graphs

There are many challenges involved in building near-term large-scale quantum computers. Some of these challenges can be overcome by partitioning a quantum circuit into smaller parts and allowing each part to be executed on a smaller quantum unit. This approach is known as distributed quantum computation. In this study, a dynamic programming algorithm is proposed to minimize the number of communications in a distributed quantum circuit. This algorithm consists of two steps: first, the quantum circuit is converted into a bipartite graph model, and then a dynamic programming approach is proposed to partition the model into low-capacity quantum circuits. The proposed approach is evaluated on some benchmark quantum circuits, and a remarkable reduction in the number of required teleportations is obtained.

Using favorite data to analyze asymmetric competition: Machine learning models

User-generated data enable businesses to derive competitive intelligence from the perspective of customers. With the advent of favorite data, we propose a sparse biterm-based Dirichlet process model and bipartite graph model with a random walk algorithm to analyze asymmetric competition. Through investigating the asymmetric market structure, representativeness degree of each entity, and competition network, the proposed machine learning models provide managerial insights into market competition. Based on 832,897 customers’ favorite lists, we empirically employ the proposed models to analyze the competition among the 2,204 car models in China's automotive market. Drawing on the activities of many customers in the market, our models enhance firms’ understanding of how the market is segmented by customers, which segments are most popular, and how entities compete with each other within the competition network. We can also inform managers of the segments within which a product competes, the representativeness of its competitors, and the leaders in each segment. The empirical results demonstrate that our models are practical in deriving insights about asymmetric competition for managers.

Collective Avoidance of Social and Health Venues and HIV Racial Inequities: Network Modeling of Venue Avoidance on Venue Affiliation, Social Networks, and HIV Risk.

Many younger Black men who have sex with men (YBMSM) are exposed to homonegativity, societal stigma, and racial discrimination in their social environment. This study uses a social network modeling methodology to identify aspects of the social environment that are not often described, that is, the places and spaces or "venues" where YBMSM socialize or where they receive HIV prevention services. In particular, we identify the structural features of avoidance of these venues as an indicator of negative experiences, using bipartite exponential random graph models. Our study theorizes that YBMSM avoid certain venues en masse through information diffusion among social network members. We specify two social mechanisms of collective venue avoidance-(1) homophily (i.e., ego-alter similarity in venue avoidance) and (2) popular opinion leaders (as early adopters)-and test the corresponding hypotheses that (Hypothesis 1) socially connected individuals avoid venues together and that (Hypothesis 2) popular individuals would be more likely to avoid venues. Based on data collected from YBMSM aged 16 to 29 years between 2014 and 2016 in Houston, Texas (N = 227) and Chicago, Illinois (N = 241), results indicate that Hypothesis 1 was supported in both cities but that Hypothesis 2 was supported only in Chicago. The findings suggest that the structural patterns of venue avoidance are different between cities and may inform dissemination of prevention messages and delivery of venue- and social influence-based HIV interventions.

Optimizing Cost and Quality of Innovative Solutions at Design Stages of Life Cycle of Engineering High-Tech Items

The article develops a method for optimizing the cost and quality of innovative solutions at design stages of a life cycle of engineering high-tech items. The method allows evaluating and comparing a whole set of permissible options of innovative solutions on the basis of all partial and generalized criteria at the stage of production design engineering.The method is based upon a bipartite graph model of component interconnections in engineering high-tech items and the criteria of optimal option selection. To evaluate the options of innovative solutions, the authors form a set of partial criteria united in the categories: production criteria; investment criteria; criteria of evaluating the quality of manufactured products while using a high-tech item; economic criteria; financial criteria; criteria of evaluation on risk types; social criteria; environmental criteria; budget criteria; integral criteria evaluating the quality of an engineering high-tech item on the basis of all indices. To evaluate the cost and quality of innovative solutions, the authors form integrated efficiency functions presented as additive convolutions of partial criteria.The developed method is automated with the use of software tools MS Excel, MS Access, C#. The method industrial approbation demonstrates its operation efficiency and universal nature for solving various problems of the comparative evaluation of innovative solutions at the design stages of a life cycle of engineering high-tech items.

A multi-feature bipartite graph ensemble for image segmentation

Image segmentation benefits from using multi-feature ensembles. In this paper, we propose a novel multi-layer bipartite graph model for more effective feature fusion. This model employs multiple graph layers, each representing a feature space. They share common vertices but have individual edge sets that are obtained from different feature spaces. The features are fused by the regularization defined on a Grassmann manifold, which compresses the graph layers into subspace representations and merges them into one. The merged graph is then fed into a graph-cut algorithm to generate the final segmentation. Experiments carried out using the Berkeley segmentation benchmark show that our model is effective in feature fusion and image segmentation, outperforming the state-of-the-art superpixel-based methods.

Wireless Channel Induced Coding

In this letter, we present a new wireless channel induced coding (WiCiC) scheme over the binary field in order to eliminate the performance loss which occurs due to not taking the channel conditions into account while determining network code coefficients. We utilize the composite channel-state matrix composed of source-to-relay and relay-to-destination channel-state matrices, as a result of the transition from the tripartite graph model to the bipartite graph model of the network. Through the use of this composite channel-state matrix, the encoding problem can be handled as a maximum flow problem owing to the max-flow min-cut theorem. The proposed WiCiC scheme has a lower decoding complexity and decoding delay compared with the benchmark studies while achieving the optimum decoding failure performance and providing user fairness. Extensive simulation results are presented to make reasonable inferences about the decoding performance of the WiCiC scheme.

A method for building a genome-connectome bipartite graph model

It has been widely shown that genomic factors influence both risk for schizophrenia and variation in functional brain connectivity. Moreover, schizophrenia is characterized by disrupted brain connectivity. In this work, we proposed a genome-connectome bipartite graph model to perform imaging genomic analysis. Functional network connectivity (FNC) was estimated after decomposing resting state functional magnetic resonance imaging data from both healthy controls (HC) and patients with schizophrenia (SZ) into spatial brain components using group independent component analysis (G-ICA). Then 83 FNC connections showing a group difference (HC vs SZ) were selected as fMRI nodes, and eighty-one schizophrenia-related single nucleotide polymorphisms (SNPs) were selected as genetic nodes respectively in the bipartite graph. Edges connecting pairs of genetic and fMRI nodes were defined based on the SNP-FNC associations across subjects evaluated by a general linear model. Results show that some SNP nodes in the bipartite graph have a high degree implying they are influential in modulating brain connectivity and may be more strongly associated with the risk of schizophrenia than other SNPs. A bi-clustering analysis detected a cluster with 15 SNPs interacting with 38 FNC connections, most of which were within or between somato-motor and visual brain areas. This suggests that the activity of these brain regions may be related to common SNPs and provides insights into the pathology of schizophrenia. The findings suggest that the SNP-FNC bipartite graph approach is a novel model to investigate genetic influences on functional brain connectivity in mental illness.

Network Structure and Spinoff Effects in a Collaborative Public Program

Despite a growing body of literature on program implementation networks, most studies have focused on understanding the network structures involving the implementation of initial or inceptive public programs. Little attention has been paid to what network structures actors create to pursue spinoff effects. Spinoff effects are defined as offshoot projects that take place ex-post the completion of the initial public projects or programs. Framing postproject collaboration between communities (villages) and organizations for spinoff projects as social networks, this research examines the network structures that drive postproject collaboration. The network data on postproject collaboration came from a field survey of 62 communities from Nepal that were engaged in building ties with organizations to mobilize resources for spinoff projects. The results from bipartite exponential random graph models applied to the network data show that postproject collaboration ties are influenced by network centralization around communities, or a greater variance among communities in the number of ties with organizations, reflecting differences in the communities’ needs for spinoff projects. In addition, communities use network bridging, or indirect ties to other communities through the partner organizations, to access new information valuable to spinoff projects. Given the widespread use of public programs, these findings provide important insights to communities and managers as they advocate postproject collaboration to realize spinoff effects and thereby to sustain the impact of public programs.

A Computational Bipartite Graph-Based Drug Repurposing Method.

We present a bipartite graph-based approach to calculate drug pairwise similarity for identifying potential new indications of approved drugs. Both chemical and molecular features were used in drug similarity calculation. In this paper, we first extracted drug chemical structures and drug-target interactions. Second, we computed chemical structure similarity and drug- target profile similarity. Further, we constructed a bipartite graph model with known relationships between drugs and their target proteins. Finally, we weighted summing drug structure similarity with target profile similarity to derive drug pairwise similarity, so that we can predict potential indication of a drug from its similar drugs. In addition, we summarized some alternative strategies and variations follow-up to each section in the overall analysis.

Defect Tolerance: Fundamental Limits and Examples

This paper addresses the problem of adding redundancy to a collection of physical objects so that the overall system is more robust to failures. In contrast to its information counterpart, which can exploit parity to protect multiple information symbols from a single erasure, physical redundancy can only be realized through duplication and substitution of objects. We propose a bipartite graph model for designing defect-tolerant systems, in which the defective objects are replaced by the judiciously connected redundant objects. The fundamental limits of this model are characterized under various asymptotic settings and both asymptotic and finite-size systems that approach these limits are constructed. Among other results, we show that the simple modular redundancy is in general suboptimal. As we develop, this combinatorial problem of defect tolerant system design has a natural interpretation as one of graph coloring, and the analysis is significantly different from that traditionally used in information redundancy for error-control codes.