Information retrieval is useful in all aspects of life, ranging from clothing shopping to educati... more Information retrieval is useful in all aspects of life, ranging from clothing shopping to education and academic pursuits. Many systems optimize models with pairwise ranking techniques such as Bayesian Personalized Ranking (BPR) for personalized information retrieval. A Bayesian personalized ranking system can assist specific shoppers, students, and researchers based on their interaction, which has illustrated an enormous capability for improvement by generating intelligent recommendations, such as clothing, books, and other related information. However, for such users, finding the desired clothing and books online is complex and is influenced by various factors (e.g., visual appearance and time). As such, traditional personalized recommendation methods that model only user-product interaction data would deliver unsatisfactory recommendation results. In this paper, we propose combining visual, temporal, and sequential information for personalized recommendations. Technically speaking, our main contributions include: (1) We incorporate the image features of clothing and books into personalized ranking to model users’ preferences. (2) We design a new time model for personalized recommender systems. The visual features are then injected into the time model to capture the temporal dynamics of visual preferences. (3) To this end, we present a Time Hierarchical Embedding (T-Sherlock) approach, which can incorporate sequential and temporal information simultaneously to model users’ preferences for different categories of products. To reduce the impact of adversarial noise, we train a T-Sherlock objective function using minimax adversarial training (AT-Sherlock). Experiments on real-world datasets demonstrated the efficacy of our methods in comparison to baselines.
2017 International Workshop on Complex Systems and Networks (IWCSN), 2017
Recommended systems are becoming important and popular for online shopping platforms and vendors.... more Recommended systems are becoming important and popular for online shopping platforms and vendors. Moreover, one-class collaborative filtering, which is based on the modeling of the feedback records of E-commercial website consumers, is one of the most widely used recommendation algorithms both academically and practically. However, one significant drawback of the existing techniques is that they typically do nor consider ratings and visual-temporal contexts, which are useful and important in modeling user behaviors. Therefore, to address this problem, we propose a new recommendation algorithm which is based on the combination of image features, user feedback ratings, and product evolution trends. The image feature can be extracted automatically using deep convolution neural network. Thus, our technique, which is essentially a time-aware visual mode, can represent the different visual feature preference of users over time. Our model is evaluated using the widely adopted Amazon online data and shown significantly improvements.
Machine learning classifiers are susceptible to adversarial perturbations, and their existence ra... more Machine learning classifiers are susceptible to adversarial perturbations, and their existence raises security concerns with a focus on recommendation systems. While there is a substantial effort to investigate attacks and defensive techniques in recommendation systems, Basic Iterative perturbation strategies (BIM) have been under-researched in multimedia recommendation. In this work, we adapt the iterative approach for multimedia recommendation. We proposed a novel Dynamic Collaborative Filtering with Aesthetic (DCFA) approach which leverages aesthetic features of clothing images into a multi-objective pairwise ranking to capture consumer aesthetic taste at a specific time through adversarial training (ADCFA). The DCFA method extends visual recommendation to make three key contributions: (1) incorporate aesthetic features into multimedia recommender system to model consumers’ preferences in the aesthetic aspect. (2) Design a multi-objective personalized ranking for the visual recommendation. (3) Use the aesthetic features to optimize the learning strategy to capture the temporal dynamics of image aesthetic preferences. To reduce the impact of perturbation, we train a DCFA objective function using minimax adversarial training. Extensive experiments on three datasets demonstrate the effectiveness of our method.
Machine learning classifiers are vulnerable to adversarial perturbation, and their presence raise... more Machine learning classifiers are vulnerable to adversarial perturbation, and their presence raises security concerns, especially in recommendation systems. While attacks and defense mechanisms in recommendation systems have received significant attention, Basic Iterative Method (BIM), which has been shown in Computer Vision to increase attack effectiveness by more than 60%, has received little attention in ownership recommendation. As a result, ownership recommender systems may be more sensitive to iterative perturbations, resulting in significant generalization errors. Adversarial Training, a regularization strategy that can withstand worst-case iterative perturbations, could be a viable option for improving model robustness and generalization. In this paper, we implement BIM for ownership recommendations. Through adversarial training, we propose the Adversarial Consumer and Producer Recommendation (ACPR) approach that integrates ownership features into a multi-objective pairwise ranking to capture the user’s preferences. The ACPR method learns a core embedding for each user and two transformation matrices that project the user’s core embedding into two role embeddings (i.e., a producer and consumer role) using an extension of matrix factorization. To minimize the impact of iterative perturbation, we train a consumer and producer recommender objective function using minimax adversarial training. Empirical studies on two Large-scale applications show that our method outperforms standard recommendation methods and recent methods that model ownership information.
Abstract In this study, a bisphenol containing spiral ring (BP-310) was synthesized with a facile... more Abstract In this study, a bisphenol containing spiral ring (BP-310) was synthesized with a facile one-step reaction routine. It was copolymerized with terephthaloyl acid chloride (TPC)/isophthalic acid chloride (TPC) and bisphenol A (BPA) to yield intrinsic porous co-polyarylates. The resultant co-polyarylates were found to have good thermal properties with glass transition temperature of 199.6–223.3 °C and initial degradation temperature of 463–482 °C. They could be dissolved in dichloromethane, DMSO, NMP and supplied transparent films with transparency of 86.2–89.4% at 450 nm. Also they were found to have good melt-flowability with melt viscosity of 414–6756 Pas (282–310 °C) and can be injected into different samples with good tensile strength and modulus of 71–73.9 MPa and 1.8–2.4 GPa. With the introduction of intrinsic porous spiral ring, the dielectric constant of the co-polyarylates decreased gradually from 3.0 down to 2.16 (at 107 Hz), tan loss factor reduced from 4.3*10−2 to 2.3*10−2 (at 107 Hz), 9.7*10−3 to 8.3*10−3 (at 104 Hz) respectively. As we expected, with the introduction of spiral ring into the polyarylate molecular chain provides a facile and low cost routine to fabricate intrinsic porous high performance materials, it can be potentially applied in these fields such as high frequency communication system such as 5G/6G.
In this study, a novel bisphenol containing cyclohexane group (CDBMP) was synthesized. It was rea... more In this study, a novel bisphenol containing cyclohexane group (CDBMP) was synthesized. It was reacted with 4, 4'-dichlorodiphenylsulfone (DCDPS) to synthesize polyether sulfone containing-OCH 3 group (PES-OCH 3). Then PES-OCH 3 was reacted with BBr 3 and (t-Boc) 2 O to yield PES-OH and PES-Boc. The glass transition temperature of PES-OCH 3 and PES-OH was found to be 177.6 and 240.7 o C, respectively. The resultant resins exhibited good tensile strength and solubility (dissolved in NMP, DMF and so on). Interestingly, PES-Boc showed excellent foam property, the density of the foam can decreased from 1.23 to 0.012 g/cm 3. Also we found the dielectric constant and dielectric loss value (10 7 Hz) of the foaming sample
IEEE Transactions on Circuits and Systems II: Express Briefs, 2020
Deep neural networks are becoming popular and important assets of many AI companies. However, rec... more Deep neural networks are becoming popular and important assets of many AI companies. However, recent studies indicate that they are also vulnerable to adversarial attacks. Adversarial attacks can be either white-box or black-box. The white-box attacks assume full knowledge of the models while the black-box ones assume none. In general, revealing more internal information can enable much more powerful and efficient attacks. However, in most real-world applications, the internal information of embedded AI devices is unavailable, i.e., they are black-box. Therefore, in this work, we propose a side-channel information based technique to reveal the internal information of black-box models. Specifically, we have made the following contributions: (1) we are the first to use side-channel information to reveal internal network architecture in embedded devices; (2) we are the first to construct models for internal parameter estimation; and (3) we validate our methods on real-world devices and applications. The experimental results show that our method can achieve 96.50% accuracy on average. Such results suggest that we should pay strong attention to the security problem of many AI applications, and further propose corresponding defensive strategies in the future.
IEEE Transactions on Systems, Man, and Cybernetics: Systems, 2019
Nonintrusive orientation detection is an important yet largely unaddressed area. It can be used i... more Nonintrusive orientation detection is an important yet largely unaddressed area. It can be used in many important applications, e.g., interactive games, medical care, and various industrial scenarios. Moreover, our novel techniques can be readily implemented in other critical areas, such as indoor localization, objective tracking, movement detection, etc. Many factors can limit the performance of detection algorithms in realword applications, an important one of which is the negligence of subcarrier correlations. Therefore, we propose to build our system based on existing WiFi infrastructure and its channel state information (CSI). To explore the correlations of adjacent subcarriers, we apply the visibility graph (VG)-based network analysis method to process the CSI data. Specifically, in this article we make the following contributions: 1) we design a CSIbased orientation detection system; 2) we model the correlations between subcarriers using complex network and propose a VGbased feature extraction technique; and 3) we demonstrate the performance and effectiveness of our system with commercial products in real-world deployments. The experimental results show that our technique can achieve more than 98% accuracy and at least 26% better than the baseline approaches.
International Journal of Software Engineering and Knowledge Engineering, 2018
Open source software (OSS) projects and communities are becoming increasingly popular and influen... more Open source software (OSS) projects and communities are becoming increasingly popular and influential recently. Communications and collaborations are essential for the success of projects. Usually, the most active and productive programmers are awarded with promotion to developers. To more effectively manage and progress the projects, it is important and beneficial to rank the programmers and thus, predict the developer candidates. In this work, we propose to combine machine learning techniques with existing complex network node ranking algorithms to improve the prediction results. Specifically, we have made the following contributions: (1), we have designed a novel machine learning-based classifier with significantly improved prediction performance; (2), we have constructed and tested various networks built based on the programmer email communication information; and (3), we have used real-world project data to compare different techniques and validate our methods. Experimental res...
IEEE Transactions on Computational Social Systems, 2018
Animals search for food based on certain optimal principles and over time form foraging patterns ... more Animals search for food based on certain optimal principles and over time form foraging patterns effective for survival in changing environments. Due to the many choices available in modern society, we also face a decision on where to get their food. We call this "modern human food foraging," since the Internet makes foraging much more convenient than before. People search online for food venues, or restaurants, through websites such as Yelp, and write reviews for the food they tasted, which in turn, facilitate others' searches in the future. These activities make the whole community of restaurant patrons wiser over time. Moreover, the archives of all these choices and evaluations are publicly available, and can help researchers better understand human foraging patterns in modern society. In this paper, we use a Yelp data set to study modern human food foraging patterns, with respect to both geography and cuisine. To understand spatial patterns, we cluster reviewed restaurants geographically and construct a taste similarity network, representing the topology of restaurant cuisine space. We find that people steadily expand their foraging domains from the nearest to them to the distant in geography and from the most familiar to the novel in cuisine. Using longitudinal data of restaurant reviews, we build a geographical foraging network and a taste foraging network for each patron based on which, we propose three kinds of entropies to characterize foraging patterns. We show that the modern foraging patterns of restaurant patrons in both geography and cuisine are of high regularity, indicating that their behaviors are rather predictable. The foraging patterns are also associated with individual social status in the community. Namely, people having a higher variety in the restaurant cuisines they have visited, but fewer actual locations they visited, tend to attract more followers.
Industrial & Engineering Chemistry Research, 2019
A series of thermoplastic and optical transparent polyimide bearing aliphatic ring was manufactur... more A series of thermoplastic and optical transparent polyimide bearing aliphatic ring was manufactured through a two-step approach. The solubility and melt processability were significantly enhanced after embedded with flexible thio-ether moieties. The complex viscosity was within 225.6-3980 Pa•s from 330 °C to 300 °C. The modified polyimide exhibited moderate mechanical strength to 76 MPa. The 5% thermal weight loss temperature was around 443 °C, with glass transition temperature of 204.6-234.8 °C. Besides, the cyclic aliphatic segments endowed the obtained PIs with improved optical performance, the transmittance could be up to 82% at 450 nm, while the transmittance of industrialized thermoplastic polyimide derived from PMDA and ODA was less than 20% under the same condition.
IEEE Transactions on Knowledge and Data Engineering, 2018
Real-world networks feature weights of interactions, where link weights often represent some phys... more Real-world networks feature weights of interactions, where link weights often represent some physical attributes. In many situations, to recover the missing data or predict the network evolution, we need to predict link weights in a network. In this paper, we first proposed a series of new centrality indices for links in line graph. Then, utilizing these line graph indices, as well as a number of original graph indices, we designed three supervised learning methods to realize link weight prediction both in the networks of single layer and multiple layers, which perform much better than several recently proposed baseline methods. We found that the resource allocation index (RA) plays a more important role in the weight prediction than other topological properties, and the line graph indices are at least as important as the original graph indices in link weight prediction. In particular, the success application of our methods on Yelp layered network suggests that we can indeed predict the offline co-foraging behaviors of users just based on their online social interactions, which may open a new direction for link weight prediction algorithms, and meanwhile provide insights to design better restaurant recommendation systems.
Passive indoor localization techniques can have many important applications. They are nonintrusiv... more Passive indoor localization techniques can have many important applications. They are nonintrusive and do not require users carrying measuring devices. Therefore, indoor localization techniques are widely used in many critical areas, such as security, logistics, healthcare, etc. However, because of the unpredictable indoor environment dynamics, the existing nonintrusive indoor localization techniques can be quite inaccurate, which greatly limits their real-world applications. To address those problems, in this work, we develop a channel state information (CSI) based indoor localization technique. Unlike the existing methods, we employ both the intra-subcarrier statistics features and the inter-subcarrier network features. Specifically, we make the following contributions: (1) we design a novel passive indoor localization algorithm which combines the statistics and network features; (2) we modify the visibility graph (VG) technique to build complex networks for the indoor localizatio...
IEEE Transactions on Systems, Man, and Cybernetics: Systems, 2017
Passive indoor localization is important. Unlike active localization techniques, it does not requ... more Passive indoor localization is important. Unlike active localization techniques, it does not require for users to carry measuring devices, e.g., smart phones. Thus, it is widely used in applications such as security, smart housing, object tracking, etc. However, in real-world applications, the passive localization accuracy is limited due to the environment noises, multipath effect, etc. To address those problems, in this paper, we propose to use channel state information (CSI) instead. Specifically, we make the following contributions: 1) we design a CSI-based passive indoor localization system; 2) we develop a Naive Bayes classifier enhanced with confidence level information; and 3) we demonstrate the effectiveness of our technique using real-world deployments. The experimental results show that our technique can achieve more than 86% accuracy on average and at least 15% better than the baseline Naive Bayes classifier.
Physica A: Statistical Mechanics and its Applications, 2017
In pinning control of complex networks, it is found that, with the same pinning effort, the netwo... more In pinning control of complex networks, it is found that, with the same pinning effort, the network can be better controlled by pinning the large-degree nodes. But in the clustered complex networks, this preferential pinning (PP) strategy is losing its effectiveness. In this paper, we demonstrate that in the clustered complex networks, especially when the clusters have different size, the random pinning (RP) strategy performs much better than the PP strategy. Then, we propose a new pinning strategy based on cluster degree. It is revealed that the new cluster pinning strategy behaves better than RP strategy when there are only a smaller number of pinning nodes. The mechanism is studied by using eigenvalue and eigenvector analysis, and the simulations of coupled chaotic oscillators are given to verify the theoretical results. These findings could be beneficial for the design of control schemes in some practical systems.
Optik - International Journal for Light and Electron Optics, 2016
Abstract A broadband photonic crystal (PhC) waveguide crossing is proposed with low crosstalk. On... more Abstract A broadband photonic crystal (PhC) waveguide crossing is proposed with low crosstalk. On the basis of the structure, two 1 × 2 PhC λ-routers are constructed by exploiting several PhC point-defect micro-cavities. By means of the interference coupling between the point-defect modes in micro-cavities, the incidence light with the resonant wavelength can be routed to the designed output port in the router. To achieve the high routing efficiency with narrow pass-band width, the micro-cavity structure is elaborately engineered to satisfy the condition of the quality factor ratio, which consists of these rods with the gradual radii. Their performances are demonstrated by the numerical results calculated by the finite difference time domain (FDTD) and the plane wave expansion (PWE) methods. Furthermore, the analysis of the 4 × 4 λ -router configuration, obtained by assembling these basic wavelength routing elements, is reported to highlight its performances with a theoretical maximum crosstalk between the ports equal to −27.23 dB.
Information retrieval is useful in all aspects of life, ranging from clothing shopping to educati... more Information retrieval is useful in all aspects of life, ranging from clothing shopping to education and academic pursuits. Many systems optimize models with pairwise ranking techniques such as Bayesian Personalized Ranking (BPR) for personalized information retrieval. A Bayesian personalized ranking system can assist specific shoppers, students, and researchers based on their interaction, which has illustrated an enormous capability for improvement by generating intelligent recommendations, such as clothing, books, and other related information. However, for such users, finding the desired clothing and books online is complex and is influenced by various factors (e.g., visual appearance and time). As such, traditional personalized recommendation methods that model only user-product interaction data would deliver unsatisfactory recommendation results. In this paper, we propose combining visual, temporal, and sequential information for personalized recommendations. Technically speaking, our main contributions include: (1) We incorporate the image features of clothing and books into personalized ranking to model users’ preferences. (2) We design a new time model for personalized recommender systems. The visual features are then injected into the time model to capture the temporal dynamics of visual preferences. (3) To this end, we present a Time Hierarchical Embedding (T-Sherlock) approach, which can incorporate sequential and temporal information simultaneously to model users’ preferences for different categories of products. To reduce the impact of adversarial noise, we train a T-Sherlock objective function using minimax adversarial training (AT-Sherlock). Experiments on real-world datasets demonstrated the efficacy of our methods in comparison to baselines.
2017 International Workshop on Complex Systems and Networks (IWCSN), 2017
Recommended systems are becoming important and popular for online shopping platforms and vendors.... more Recommended systems are becoming important and popular for online shopping platforms and vendors. Moreover, one-class collaborative filtering, which is based on the modeling of the feedback records of E-commercial website consumers, is one of the most widely used recommendation algorithms both academically and practically. However, one significant drawback of the existing techniques is that they typically do nor consider ratings and visual-temporal contexts, which are useful and important in modeling user behaviors. Therefore, to address this problem, we propose a new recommendation algorithm which is based on the combination of image features, user feedback ratings, and product evolution trends. The image feature can be extracted automatically using deep convolution neural network. Thus, our technique, which is essentially a time-aware visual mode, can represent the different visual feature preference of users over time. Our model is evaluated using the widely adopted Amazon online data and shown significantly improvements.
Machine learning classifiers are susceptible to adversarial perturbations, and their existence ra... more Machine learning classifiers are susceptible to adversarial perturbations, and their existence raises security concerns with a focus on recommendation systems. While there is a substantial effort to investigate attacks and defensive techniques in recommendation systems, Basic Iterative perturbation strategies (BIM) have been under-researched in multimedia recommendation. In this work, we adapt the iterative approach for multimedia recommendation. We proposed a novel Dynamic Collaborative Filtering with Aesthetic (DCFA) approach which leverages aesthetic features of clothing images into a multi-objective pairwise ranking to capture consumer aesthetic taste at a specific time through adversarial training (ADCFA). The DCFA method extends visual recommendation to make three key contributions: (1) incorporate aesthetic features into multimedia recommender system to model consumers’ preferences in the aesthetic aspect. (2) Design a multi-objective personalized ranking for the visual recommendation. (3) Use the aesthetic features to optimize the learning strategy to capture the temporal dynamics of image aesthetic preferences. To reduce the impact of perturbation, we train a DCFA objective function using minimax adversarial training. Extensive experiments on three datasets demonstrate the effectiveness of our method.
Machine learning classifiers are vulnerable to adversarial perturbation, and their presence raise... more Machine learning classifiers are vulnerable to adversarial perturbation, and their presence raises security concerns, especially in recommendation systems. While attacks and defense mechanisms in recommendation systems have received significant attention, Basic Iterative Method (BIM), which has been shown in Computer Vision to increase attack effectiveness by more than 60%, has received little attention in ownership recommendation. As a result, ownership recommender systems may be more sensitive to iterative perturbations, resulting in significant generalization errors. Adversarial Training, a regularization strategy that can withstand worst-case iterative perturbations, could be a viable option for improving model robustness and generalization. In this paper, we implement BIM for ownership recommendations. Through adversarial training, we propose the Adversarial Consumer and Producer Recommendation (ACPR) approach that integrates ownership features into a multi-objective pairwise ranking to capture the user’s preferences. The ACPR method learns a core embedding for each user and two transformation matrices that project the user’s core embedding into two role embeddings (i.e., a producer and consumer role) using an extension of matrix factorization. To minimize the impact of iterative perturbation, we train a consumer and producer recommender objective function using minimax adversarial training. Empirical studies on two Large-scale applications show that our method outperforms standard recommendation methods and recent methods that model ownership information.
Abstract In this study, a bisphenol containing spiral ring (BP-310) was synthesized with a facile... more Abstract In this study, a bisphenol containing spiral ring (BP-310) was synthesized with a facile one-step reaction routine. It was copolymerized with terephthaloyl acid chloride (TPC)/isophthalic acid chloride (TPC) and bisphenol A (BPA) to yield intrinsic porous co-polyarylates. The resultant co-polyarylates were found to have good thermal properties with glass transition temperature of 199.6–223.3 °C and initial degradation temperature of 463–482 °C. They could be dissolved in dichloromethane, DMSO, NMP and supplied transparent films with transparency of 86.2–89.4% at 450 nm. Also they were found to have good melt-flowability with melt viscosity of 414–6756 Pas (282–310 °C) and can be injected into different samples with good tensile strength and modulus of 71–73.9 MPa and 1.8–2.4 GPa. With the introduction of intrinsic porous spiral ring, the dielectric constant of the co-polyarylates decreased gradually from 3.0 down to 2.16 (at 107 Hz), tan loss factor reduced from 4.3*10−2 to 2.3*10−2 (at 107 Hz), 9.7*10−3 to 8.3*10−3 (at 104 Hz) respectively. As we expected, with the introduction of spiral ring into the polyarylate molecular chain provides a facile and low cost routine to fabricate intrinsic porous high performance materials, it can be potentially applied in these fields such as high frequency communication system such as 5G/6G.
In this study, a novel bisphenol containing cyclohexane group (CDBMP) was synthesized. It was rea... more In this study, a novel bisphenol containing cyclohexane group (CDBMP) was synthesized. It was reacted with 4, 4'-dichlorodiphenylsulfone (DCDPS) to synthesize polyether sulfone containing-OCH 3 group (PES-OCH 3). Then PES-OCH 3 was reacted with BBr 3 and (t-Boc) 2 O to yield PES-OH and PES-Boc. The glass transition temperature of PES-OCH 3 and PES-OH was found to be 177.6 and 240.7 o C, respectively. The resultant resins exhibited good tensile strength and solubility (dissolved in NMP, DMF and so on). Interestingly, PES-Boc showed excellent foam property, the density of the foam can decreased from 1.23 to 0.012 g/cm 3. Also we found the dielectric constant and dielectric loss value (10 7 Hz) of the foaming sample
IEEE Transactions on Circuits and Systems II: Express Briefs, 2020
Deep neural networks are becoming popular and important assets of many AI companies. However, rec... more Deep neural networks are becoming popular and important assets of many AI companies. However, recent studies indicate that they are also vulnerable to adversarial attacks. Adversarial attacks can be either white-box or black-box. The white-box attacks assume full knowledge of the models while the black-box ones assume none. In general, revealing more internal information can enable much more powerful and efficient attacks. However, in most real-world applications, the internal information of embedded AI devices is unavailable, i.e., they are black-box. Therefore, in this work, we propose a side-channel information based technique to reveal the internal information of black-box models. Specifically, we have made the following contributions: (1) we are the first to use side-channel information to reveal internal network architecture in embedded devices; (2) we are the first to construct models for internal parameter estimation; and (3) we validate our methods on real-world devices and applications. The experimental results show that our method can achieve 96.50% accuracy on average. Such results suggest that we should pay strong attention to the security problem of many AI applications, and further propose corresponding defensive strategies in the future.
IEEE Transactions on Systems, Man, and Cybernetics: Systems, 2019
Nonintrusive orientation detection is an important yet largely unaddressed area. It can be used i... more Nonintrusive orientation detection is an important yet largely unaddressed area. It can be used in many important applications, e.g., interactive games, medical care, and various industrial scenarios. Moreover, our novel techniques can be readily implemented in other critical areas, such as indoor localization, objective tracking, movement detection, etc. Many factors can limit the performance of detection algorithms in realword applications, an important one of which is the negligence of subcarrier correlations. Therefore, we propose to build our system based on existing WiFi infrastructure and its channel state information (CSI). To explore the correlations of adjacent subcarriers, we apply the visibility graph (VG)-based network analysis method to process the CSI data. Specifically, in this article we make the following contributions: 1) we design a CSIbased orientation detection system; 2) we model the correlations between subcarriers using complex network and propose a VGbased feature extraction technique; and 3) we demonstrate the performance and effectiveness of our system with commercial products in real-world deployments. The experimental results show that our technique can achieve more than 98% accuracy and at least 26% better than the baseline approaches.
International Journal of Software Engineering and Knowledge Engineering, 2018
Open source software (OSS) projects and communities are becoming increasingly popular and influen... more Open source software (OSS) projects and communities are becoming increasingly popular and influential recently. Communications and collaborations are essential for the success of projects. Usually, the most active and productive programmers are awarded with promotion to developers. To more effectively manage and progress the projects, it is important and beneficial to rank the programmers and thus, predict the developer candidates. In this work, we propose to combine machine learning techniques with existing complex network node ranking algorithms to improve the prediction results. Specifically, we have made the following contributions: (1), we have designed a novel machine learning-based classifier with significantly improved prediction performance; (2), we have constructed and tested various networks built based on the programmer email communication information; and (3), we have used real-world project data to compare different techniques and validate our methods. Experimental res...
IEEE Transactions on Computational Social Systems, 2018
Animals search for food based on certain optimal principles and over time form foraging patterns ... more Animals search for food based on certain optimal principles and over time form foraging patterns effective for survival in changing environments. Due to the many choices available in modern society, we also face a decision on where to get their food. We call this "modern human food foraging," since the Internet makes foraging much more convenient than before. People search online for food venues, or restaurants, through websites such as Yelp, and write reviews for the food they tasted, which in turn, facilitate others' searches in the future. These activities make the whole community of restaurant patrons wiser over time. Moreover, the archives of all these choices and evaluations are publicly available, and can help researchers better understand human foraging patterns in modern society. In this paper, we use a Yelp data set to study modern human food foraging patterns, with respect to both geography and cuisine. To understand spatial patterns, we cluster reviewed restaurants geographically and construct a taste similarity network, representing the topology of restaurant cuisine space. We find that people steadily expand their foraging domains from the nearest to them to the distant in geography and from the most familiar to the novel in cuisine. Using longitudinal data of restaurant reviews, we build a geographical foraging network and a taste foraging network for each patron based on which, we propose three kinds of entropies to characterize foraging patterns. We show that the modern foraging patterns of restaurant patrons in both geography and cuisine are of high regularity, indicating that their behaviors are rather predictable. The foraging patterns are also associated with individual social status in the community. Namely, people having a higher variety in the restaurant cuisines they have visited, but fewer actual locations they visited, tend to attract more followers.
Industrial & Engineering Chemistry Research, 2019
A series of thermoplastic and optical transparent polyimide bearing aliphatic ring was manufactur... more A series of thermoplastic and optical transparent polyimide bearing aliphatic ring was manufactured through a two-step approach. The solubility and melt processability were significantly enhanced after embedded with flexible thio-ether moieties. The complex viscosity was within 225.6-3980 Pa•s from 330 °C to 300 °C. The modified polyimide exhibited moderate mechanical strength to 76 MPa. The 5% thermal weight loss temperature was around 443 °C, with glass transition temperature of 204.6-234.8 °C. Besides, the cyclic aliphatic segments endowed the obtained PIs with improved optical performance, the transmittance could be up to 82% at 450 nm, while the transmittance of industrialized thermoplastic polyimide derived from PMDA and ODA was less than 20% under the same condition.
IEEE Transactions on Knowledge and Data Engineering, 2018
Real-world networks feature weights of interactions, where link weights often represent some phys... more Real-world networks feature weights of interactions, where link weights often represent some physical attributes. In many situations, to recover the missing data or predict the network evolution, we need to predict link weights in a network. In this paper, we first proposed a series of new centrality indices for links in line graph. Then, utilizing these line graph indices, as well as a number of original graph indices, we designed three supervised learning methods to realize link weight prediction both in the networks of single layer and multiple layers, which perform much better than several recently proposed baseline methods. We found that the resource allocation index (RA) plays a more important role in the weight prediction than other topological properties, and the line graph indices are at least as important as the original graph indices in link weight prediction. In particular, the success application of our methods on Yelp layered network suggests that we can indeed predict the offline co-foraging behaviors of users just based on their online social interactions, which may open a new direction for link weight prediction algorithms, and meanwhile provide insights to design better restaurant recommendation systems.
Passive indoor localization techniques can have many important applications. They are nonintrusiv... more Passive indoor localization techniques can have many important applications. They are nonintrusive and do not require users carrying measuring devices. Therefore, indoor localization techniques are widely used in many critical areas, such as security, logistics, healthcare, etc. However, because of the unpredictable indoor environment dynamics, the existing nonintrusive indoor localization techniques can be quite inaccurate, which greatly limits their real-world applications. To address those problems, in this work, we develop a channel state information (CSI) based indoor localization technique. Unlike the existing methods, we employ both the intra-subcarrier statistics features and the inter-subcarrier network features. Specifically, we make the following contributions: (1) we design a novel passive indoor localization algorithm which combines the statistics and network features; (2) we modify the visibility graph (VG) technique to build complex networks for the indoor localizatio...
IEEE Transactions on Systems, Man, and Cybernetics: Systems, 2017
Passive indoor localization is important. Unlike active localization techniques, it does not requ... more Passive indoor localization is important. Unlike active localization techniques, it does not require for users to carry measuring devices, e.g., smart phones. Thus, it is widely used in applications such as security, smart housing, object tracking, etc. However, in real-world applications, the passive localization accuracy is limited due to the environment noises, multipath effect, etc. To address those problems, in this paper, we propose to use channel state information (CSI) instead. Specifically, we make the following contributions: 1) we design a CSI-based passive indoor localization system; 2) we develop a Naive Bayes classifier enhanced with confidence level information; and 3) we demonstrate the effectiveness of our technique using real-world deployments. The experimental results show that our technique can achieve more than 86% accuracy on average and at least 15% better than the baseline Naive Bayes classifier.
Physica A: Statistical Mechanics and its Applications, 2017
In pinning control of complex networks, it is found that, with the same pinning effort, the netwo... more In pinning control of complex networks, it is found that, with the same pinning effort, the network can be better controlled by pinning the large-degree nodes. But in the clustered complex networks, this preferential pinning (PP) strategy is losing its effectiveness. In this paper, we demonstrate that in the clustered complex networks, especially when the clusters have different size, the random pinning (RP) strategy performs much better than the PP strategy. Then, we propose a new pinning strategy based on cluster degree. It is revealed that the new cluster pinning strategy behaves better than RP strategy when there are only a smaller number of pinning nodes. The mechanism is studied by using eigenvalue and eigenvector analysis, and the simulations of coupled chaotic oscillators are given to verify the theoretical results. These findings could be beneficial for the design of control schemes in some practical systems.
Optik - International Journal for Light and Electron Optics, 2016
Abstract A broadband photonic crystal (PhC) waveguide crossing is proposed with low crosstalk. On... more Abstract A broadband photonic crystal (PhC) waveguide crossing is proposed with low crosstalk. On the basis of the structure, two 1 × 2 PhC λ-routers are constructed by exploiting several PhC point-defect micro-cavities. By means of the interference coupling between the point-defect modes in micro-cavities, the incidence light with the resonant wavelength can be routed to the designed output port in the router. To achieve the high routing efficiency with narrow pass-band width, the micro-cavity structure is elaborately engineered to satisfy the condition of the quality factor ratio, which consists of these rods with the gradual radii. Their performances are demonstrated by the numerical results calculated by the finite difference time domain (FDTD) and the plane wave expansion (PWE) methods. Furthermore, the analysis of the 4 × 4 λ -router configuration, obtained by assembling these basic wavelength routing elements, is reported to highlight its performances with a theoretical maximum crosstalk between the ports equal to −27.23 dB.
Uploads
Papers by Zhefu Wu