In social networks,user attention affects the user’s decision-making,resulting in a performance alteration of the recommendation systems.Existing systems make recommendations mainly according to users’preferences wi...In social networks,user attention affects the user’s decision-making,resulting in a performance alteration of the recommendation systems.Existing systems make recommendations mainly according to users’preferences with a particular focus on items.However,the significance of users’attention and the difference in the influence of different users and items are often ignored.Thus,this paper proposes an attention-based multi-layer friend recommendation model to mitigate information overload in social networks.We first constructed the basic user and item matrix via convolutional neural networks(CNN).Then,we obtained user preferences by using the relationships between users and items,which were later inputted into our model to learn the preferences between friends.The error performance of the proposed method was compared with the traditional solutions based on collaborative filtering.A comprehensive performance evaluation was also conducted using large-scale real-world datasets collected from three popular location-based social networks.The experimental results revealed that our proposal outperforms the traditional methods in terms of recommendation performance.展开更多
Benefited from the design of separating control plane and data plane,software defined networking(SDN)is widely concerned and applied.Its quick response capability to network events with changes in network policies ena...Benefited from the design of separating control plane and data plane,software defined networking(SDN)is widely concerned and applied.Its quick response capability to network events with changes in network policies enables more dynamic management of data center networks.Although the SDN controller architecture is increasingly optimized for swift policy updates,the data plane,especially the prevailing ternary content-addressable memory(TCAM)based flow tables on physical SDN switches,remains unoptimized for fast rule updates,and is gradually becoming the primary bottleneck along the policy update pipeline.In this paper,we present RuleTris,the first SDN update optimization framework that minimizes rule update latency for TCAM-based switches.RuleTris employs the dependency graph(DAG)as the key abstraction to minimize the update latency.RuleTris efficiently obtains the DAGs with novel dependency preserving algorithms that incrementally build rule dependency along with the compilation process.Then,in the guidance of the DAG,RuleTris calculates the TCAM update schedules that minimize TCAM entry moves,which are themain cause of TCAM update inefficiency.In evaluation,RuleTris achieves a median of<12 ms and 90-percentile of<15ms the end-to-end perrule update latency on our hardware prototype,outperforming the state-of-the-art composition compiler CoVisor by～20 times.展开更多
Pixel-wise Code Exposure (PCE) camera is one type of compressive sensing camera that has low power consumption and high compression ratio. Moreover, a PCE camera can control individual pixel exposure time that can ena...Pixel-wise Code Exposure (PCE) camera is one type of compressive sensing camera that has low power consumption and high compression ratio. Moreover, a PCE camera can control individual pixel exposure time that can enable high dynamic range. Conventional approaches of using PCE camera involve a time consuming and lossy process to reconstruct the original frames and then use those frames for target tracking and classification. In this paper, we present a deep learning approach that directly performs target tracking and classification in the compressive measurement domain without any frame reconstruction. Our approach has two parts: tracking and classification. The tracking has been done using YOLO (You Only Look Once) and the classification is achieved using Residual Network (ResNet). Extensive experiments using mid-wave infrared (MWIR) and long-wave infrared (LWIR) videos demonstrated the efficacy of our proposed approach.展开更多
Artificial intelligence(AI)methods have become a focus of intense interest within the eye care community.This parallels a wider interest in AI,which has started impacting many facets of society.However,understanding a...Artificial intelligence(AI)methods have become a focus of intense interest within the eye care community.This parallels a wider interest in AI,which has started impacting many facets of society.However,understanding across the community has not kept pace with technical developments.What is AI,and how does it relate to other terms like machine learning or deep learning?How is AI currently used within eye care,and how might it be used in the future?This review paper provides an overview of these concepts for eye care specialists.We explain core concepts in AI,describe how these methods have been applied in ophthalmology,and consider future directions and challenges.We walk through the steps needed to develop an AI system for eye disease,and discuss the challenges in validating and deploying such technology.We argue that among medical fields,ophthalmology may be uniquely positioned to benefit from the thoughtful deployment of AI to improve patient care.展开更多
This book focuses on the relationship between theory and applications of various optimization problems in computer engineering. In the first half of the book the theoretical foundations are presented, such as some sel...This book focuses on the relationship between theory and applications of various optimization problems in computer engineering. In the first half of the book the theoretical foundations are presented, such as some selected graph algorithms, integer linear programming and complexity theory. The second half of the book brings the theory closer to the reader by applying them to real-world optimization problems. Its aim is to bridge the often significant gap between theory and applications bringing additional value to both: the theory becomes more interesting in light of a possible application and understanding the hardness and possible solutions of the real-world problem definitely benefits from a strong theoretical background. Chapter 8 is a good example of the above. Here the authors present several versions of the frequency assignment problem (FAP), which is an important practical optimization problem arising in wireless network design. It is shown how FAP can be reduced to the earlier presented graph coloring problem. It is interesting to note that often the practical problem needs significant simplification in order to fit into the model that the theory is able to handle, or the theoretical problem needs to be extended to be able to model the needs of the practical application. Various generalizations of the simple graph coloring problem such as list coloring and T-coloring are introduced to model the constraints of the FAP. With this reduction the specific engineering problem can be han-dled through well-understood mathematical models. Besides showing the reduction to the graph coloring problem, the authors apply a graph coloring solver on industry benchmark FAP instances to further understand the characteristics of the real-world FAP. They show that there are significant differences in the difficulty of the problem on random and real-world graphs and that the parameters of the particular instance play a crucial role in the hardness of the problem. They show that the FAPs show a phase transition property in every input parameter, ie. there is a critical parameter combination where the problem gets extremely hard, but otherwise the problem can be solved relatively easily even on large real-world networks. Readers will surely benefit from the unique nature of the book that brings theory and applications close together in a well-understandable yet theoretically solid way.展开更多
文摘In social networks,user attention affects the user’s decision-making,resulting in a performance alteration of the recommendation systems.Existing systems make recommendations mainly according to users’preferences with a particular focus on items.However,the significance of users’attention and the difference in the influence of different users and items are often ignored.Thus,this paper proposes an attention-based multi-layer friend recommendation model to mitigate information overload in social networks.We first constructed the basic user and item matrix via convolutional neural networks(CNN).Then,we obtained user preferences by using the relationships between users and items,which were later inputted into our model to learn the preferences between friends.The error performance of the proposed method was compared with the traditional solutions based on collaborative filtering.A comprehensive performance evaluation was also conducted using large-scale real-world datasets collected from three popular location-based social networks.The experimental results revealed that our proposal outperforms the traditional methods in terms of recommendation performance.
基金supported by National Key R&D Program of China under Grant No.2017YFB0801703the Key Research and Development Program of Zhejiang Province under Grant No.2018C01088
文摘Benefited from the design of separating control plane and data plane,software defined networking(SDN)is widely concerned and applied.Its quick response capability to network events with changes in network policies enables more dynamic management of data center networks.Although the SDN controller architecture is increasingly optimized for swift policy updates,the data plane,especially the prevailing ternary content-addressable memory(TCAM)based flow tables on physical SDN switches,remains unoptimized for fast rule updates,and is gradually becoming the primary bottleneck along the policy update pipeline.In this paper,we present RuleTris,the first SDN update optimization framework that minimizes rule update latency for TCAM-based switches.RuleTris employs the dependency graph(DAG)as the key abstraction to minimize the update latency.RuleTris efficiently obtains the DAGs with novel dependency preserving algorithms that incrementally build rule dependency along with the compilation process.Then,in the guidance of the DAG,RuleTris calculates the TCAM update schedules that minimize TCAM entry moves,which are themain cause of TCAM update inefficiency.In evaluation,RuleTris achieves a median of<12 ms and 90-percentile of<15ms the end-to-end perrule update latency on our hardware prototype,outperforming the state-of-the-art composition compiler CoVisor by～20 times.
文摘Pixel-wise Code Exposure (PCE) camera is one type of compressive sensing camera that has low power consumption and high compression ratio. Moreover, a PCE camera can control individual pixel exposure time that can enable high dynamic range. Conventional approaches of using PCE camera involve a time consuming and lossy process to reconstruct the original frames and then use those frames for target tracking and classification. In this paper, we present a deep learning approach that directly performs target tracking and classification in the compressive measurement domain without any frame reconstruction. Our approach has two parts: tracking and classification. The tracking has been done using YOLO (You Only Look Once) and the classification is achieved using Residual Network (ResNet). Extensive experiments using mid-wave infrared (MWIR) and long-wave infrared (LWIR) videos demonstrated the efficacy of our proposed approach.
文摘Artificial intelligence(AI)methods have become a focus of intense interest within the eye care community.This parallels a wider interest in AI,which has started impacting many facets of society.However,understanding across the community has not kept pace with technical developments.What is AI,and how does it relate to other terms like machine learning or deep learning?How is AI currently used within eye care,and how might it be used in the future?This review paper provides an overview of these concepts for eye care specialists.We explain core concepts in AI,describe how these methods have been applied in ophthalmology,and consider future directions and challenges.We walk through the steps needed to develop an AI system for eye disease,and discuss the challenges in validating and deploying such technology.We argue that among medical fields,ophthalmology may be uniquely positioned to benefit from the thoughtful deployment of AI to improve patient care.
文摘This book focuses on the relationship between theory and applications of various optimization problems in computer engineering. In the first half of the book the theoretical foundations are presented, such as some selected graph algorithms, integer linear programming and complexity theory. The second half of the book brings the theory closer to the reader by applying them to real-world optimization problems. Its aim is to bridge the often significant gap between theory and applications bringing additional value to both: the theory becomes more interesting in light of a possible application and understanding the hardness and possible solutions of the real-world problem definitely benefits from a strong theoretical background. Chapter 8 is a good example of the above. Here the authors present several versions of the frequency assignment problem (FAP), which is an important practical optimization problem arising in wireless network design. It is shown how FAP can be reduced to the earlier presented graph coloring problem. It is interesting to note that often the practical problem needs significant simplification in order to fit into the model that the theory is able to handle, or the theoretical problem needs to be extended to be able to model the needs of the practical application. Various generalizations of the simple graph coloring problem such as list coloring and T-coloring are introduced to model the constraints of the FAP. With this reduction the specific engineering problem can be han-dled through well-understood mathematical models. Besides showing the reduction to the graph coloring problem, the authors apply a graph coloring solver on industry benchmark FAP instances to further understand the characteristics of the real-world FAP. They show that there are significant differences in the difficulty of the problem on random and real-world graphs and that the parameters of the particular instance play a crucial role in the hardness of the problem. They show that the FAPs show a phase transition property in every input parameter, ie. there is a critical parameter combination where the problem gets extremely hard, but otherwise the problem can be solved relatively easily even on large real-world networks. Readers will surely benefit from the unique nature of the book that brings theory and applications close together in a well-understandable yet theoretically solid way.
