Redundant Graph Research Articles

Efficient filter pruning: Reducing model complexity through redundancy graph decomposition

Filter pruning has emerged as a crucial technique facilitating the efficient deployment of Convolutional Neural Networks (CNNs) on edge devices. The resultant lightweight models often present a challenging trade-off between performance and compression rate. Moreover, the excessive granularity of filter selection, coupled with intricate fine-tuning procedures, hampers the practical applicability of numerous pruning methodologies. To tackle these limitations, our paper introduces a novel pruning approach that initially constructs an undirected graph, effectively capturing convolutional layer redundancy by quantifying the similarity among output feature maps. Subsequently, our method iteratively eliminates highly redundant and low-value filters, employing both k-core decomposition and importance ranking criteria. In practical implementation, we adopt a one-shot pruning strategy, thereby minimizing the requirement for extensive fine-tuning. Comprehensive comparative experiments demonstrate the efficacy of our proposed approach in achieving remarkable improvements in model compression without compromising performance. For instance, with CIFAR-10, our method achieves state-of-the-art results on DenseNet-40 with a mere 70 fine-tuning epochs, while effectively removing 49.39% of FLOPs, resulting in a mere 0.01% accuracy loss. Furthermore, in the context of ImageNet, our approach excels by removing 48.81% of FLOPs and 23.71% of parameters from ResNet-50, exhibiting only 1.47% and 0.83% decline in Top-1 and Top-5 Accuracy respectively. The code is publicly available at https://github.com/lijiang99/redundancy-graph-decomposition.

Simulation of the engineering network in the area of one regulator

Theproposed article present problems that update the issue of mathematical modeling of the system process of designing of new and reconstruction of old engineering networks in the conditions of their perspective development. The analysis of researches in the field of designing of engineering networks taking into account uncertainty of the information is given. The main material of the article is devoted to the construction of a mathematical model of the problem of the network synthesis in the area of action of one regulator of the supply of the target product as component of the developing engineering network design process. The list of questions to be answered when designing a network is given. The problem is solved on a subgraph of the redundant graph of the general design process. It distinguishes the vertex of the regulatorof the supply of the target product, the vertices of the consumers of the target product fed by the regulator, and the permissible communications (arcs) between the consumers. Each communication is matched with its length, indicatorof reliability of functioning, characteristics of quality criteria per unit of communicationlength. Each vertex is matched with an indicatorof reliability of providing the target product. Since thequality criteria are fuzzy numbers with their membership functions, the criteria are normalized forcomparison.Corresponding indicatorfor them, together with the weighting factors, form a generalized quality criterion per unit of communicationlength. The structure of the network has a ring character, for construction of which the vertices of thegraph are divided into subset taking into account the radius of the ring. On the basis of such a graph, a basic mathematical modeling of the Boolean programming problem was built, in which generalized quality criterion are optimized taking into account the restrictions on ensuring supply reliability indicators for consumers. The decomposition process was applied to the constructed model. It allows the solution of the general problem to be reduced to the solution of theBoolean programming problem separately for each vertex in the framework of the search for optimal routes according to the constructed criterion between the previous set of vertices and the set of vertices to which the selected vertex belongs. The work proves the possibilities of such an approach. Based on the developed decomposition, an algorithm for solving the problem of the network synthesis in the area of action of one regulator of the supply of the target productis proposed.

Assessment of the Disaster Resilience of Complex Systems: The Case of the Flood Resilience of a Densely Populated City

In the last decades, resilience became officially the worldwide cornerstone to reduce the risk of disasters and improve preparedness, response, and recovery capacities. Although the concept of resilience is now clear, it is still under debate how to model and quantify it. The aim of this work was to quantify the resilience of a complex system, such as a densely populated and urbanized area, by modelling it with a graph, the mathematical representation of the system element and connections. We showed that the graph can account for the resilience characteristics included in its definition according to the United Nations General Assembly, considering two significant aspects of this definition in particular: (1) resilience is a property of a system and not of single entities and (2) resilience is a property of the system dynamic response. We proposed to represent the exposed elements of the system and their connections (i.e., the services they exchange) with a weighted and redundant graph. By mean of it, we assessed the systemic properties, such as authority and hub values and highlighted the centrality of some elements. Furthermore, we showed that after an external perturbation, such as a hazardous event, each element can dynamically adapt, and a new graph configuration is set up, taking advantage of the redundancy of the connections and the capacity of each element to supply lost services. Finally, we proposed a quantitative metric for resilience as the actual reduction of the impacts of events at different return periods when resilient properties of the system are activated. To illustrate step by step the proposed methodology and show its practical feasibility, we applied it to a pilot study: the city of Monza, a densely populated urban environment exposed to river and pluvial floods.

Методика избыточных проектных схем и метод поконтурной минимизации систем группового водоснабжения и водоотведения

Due to the uneven distribution of water resources across the territory of Russia, many settlements and even cities do not have own sources of water supply and reservoirs, where treated wastewater could be discharged. To solve this problem, group and district water supply and sewerage systems are designed, built, and developed, hundreds and even thousands of kilometers long. The construction and operation of such facilities require significant financial investments every year. Therefore, the issues of the choice of routes, the structure of facilities, and particularly the methods of transporting water and wastewater, justifying the locations of water intakes and treat treatment facilities are relevant and require special attention and development of a comprehensive methodology for optimizing the structure and parameters of water supply and sewerage systems. The paper proposes, on the basis of a previously constructed redundant graph, the method of contour minimization of the structure of facilities and ways of supplying water to end users, transporting wastewater to treatment plants by pipelines and trucks. The redundant graph includes all kinds of connections (edges, arcs) that simulate pipelines and trucks, between vertices that simulate existing and new sources of water and wastewater discharges, sewage treatment plants, pumping stations, control tanks, and end users. The method selects the best (from the point of view of minimal life cycle costs) edges, arcs, and vertices of the graph and determines the corresponding parameters of pipeline and road sections of the network and facilities. This method is based on the procedure of traversing the contours and sequentially replacing the branches of the tree with moves, while the best option is remembered in the form of a spanning tree. Based on the results of the performed numerical experiments, it was shown that the use of combined (pipeline and truck) systems for transporting water in terms of life cycle costs is optimal.

Iterative closest graph matching for non-rigid 3D/2D coronary arteries registration

Background and objective Fusion of the preoperative computed tomography angiography and intraoperative X-ray angiography images can considerably enhance the visual perception of physicians during percutaneous coronary interventions. This technique can provide 3D information of the arteries and reduce the uncertainty of 2D guidance images. For this purpose, 3D/2D vascular registration with high accuracy and robustness is crucial for performing accurate surgery.Methods In this study, we propose an iterative closest graph matching (ICGM) method that utilizes an alternative iteration framework including correspondence and transformation phases. A coarse-to-fine matching approach based on redundant graph matching is proposed for the correspondence phase. The transformation phase involves rigid and non-rigid transformations, in which rigid transformation is calculated using a closed-form solution, and non-rigid transformation is achieved using a statistical shape model established from a synthetic deformation dataset.Results The proposed method is evaluated and compared with nine state-of-the-art methods on simulated data and clinical datasets. Experiments demonstrate that our method is insensitive to the pose of data and robust to noise and deformation. Moreover, it outperforms other methods in terms of registering real data.Conclusions Given its high capture range, the proposed method can register 3D vessels without prior initialization in clinical practice.

Optimizing DNN computation graph using graph substitutions

Deep learning has achieved great success in various real-world applications. As deep neural networks (DNNs) are getting larger, the inference and training cost of DNNs increases significantly. Since one round of inference or one iteration in the training phase of a DNN is typically modeled as a computation graph, existing works propose to optimize computation graphs by performing a sequence of functionally equivalent graph substitutions, leading to higher inference and training efficiency. In this work, we formally define the Optimizing Computation Graph using Graph Substitutions (OCGGS) problem, and prove it to be NP-hard and Poly-APX-complete. We develop two exact and efficient methods to the OCGGS problem. The pruning-based algorithm eliminates the examination of redundant graph substitution sequences, and the dynamic programming with pruning algorithm makes use of the explored graph substitutions. To further speed up the search process, we propose a sampling heuristic which is effective to optimize complex computation graphs with polynomial time and space complexity. Extensive experiments on various DNN architectures and sizes are conducted to verify the effectiveness and efficiency of our proposed solutions compared with existing techniques.

Activity Organization for Friend-Making Optimization in Online Social Networks

The social presence theory in social psychology suggests that computer-mediated online interactions are inferior to face-to-face, in-person interactions. Thus, it's important to organize social activities for online social network users to meet in person. In this paper, we consider the scenarios of organizing in person friend-making social activities via online social networks (OSNs) and formulate a new research problem, namely, <i>Hop-bounded Maximum Group Friending (HMGF)</i> , that takes into consideration both existing friendships and the likelihood of new friend making in organization of the targeted in person friend-making social activities. To find a set of attendees for such social activities, HMGF is unique and challenging due to the interplay of the group size, the constraint on existing friendships, and the objective of maximizing the likelihood of friend making. We prove that HMGF is NP-Hard, and there exists no approximation algorithm for it unless <inline-formula><tex-math notation="LaTeX">$P=NP$</tex-math></inline-formula> . We also provide an Integer Linear Programming (ILP) formulation for the HMGF problem. The ILP formulation, which can be solved efficiently by a commercial solver to obtain the optimal solution for small HMGF instances, acts as a baseline approach for comparison in the evaluation of the proposed algorithm. We further propose an error-bounded approximation algorithm, <i>MaxGF</i> , to efficiently obtain the solutions very close to the optimal solutions. To boost the performance, we devise two graph-theoretical pruning strategies, namely <i>Neighbor Pruning</i> and <i>Core Pruning</i> , which can effectively avoid redundant graph explorations to improve the performance of HMGF. We also study HMGF on a class of special graphs, <i>threshold graphs</i> , which have properties very similar to many online social networks. We prove that MaxGF can obtain the optimal solution to HMGF on threshold graphs in polynomial time. We conduct a user study to validate our problem formulation and perform extensive experiments on real datasets to demonstrate the efficiency and effectiveness of our proposed algorithm. The experimental results manifest that our proposed algorithms outperform the baselines, including the ILP formulation.

Query Acceleration of Graph Databases by ID Caching Technology

In this paper, we approach the design of ID caching technology (IDCT) for graph databases, with the purpose of accelerating the queries on graph database data and avoiding redundant graph database query operations which will consume great computer resources. Traditional graph database caching technology (GDCT) needs a large memory to store data and has the problems of serious data consistency and low cache utilization. To address these issues, in the paper we propose a new technology which focuses on ID allocation mechanism and high-speed queries of ID on graph databases. Specifically, ID of the query result is cached in memory and data consistency is achieved through the real-time synchronization and cache memory adaptation. In addition, we set up complex queries and simple queries to satisfy all query requirements and design a mechanism of cache replacement based on query action time, query times, and memory capacity, thus improving the performance furthermore. Extensive experiments show the superiority of our techniques compared with the traditional query approach of graph databases.

Cooperative caching in fog radio access networks: a graph‐based approach

In this paper, cooperative caching is investigated in fog radio access networks (F-RAN). To maximize the offloaded traffic, cooperative caching optimization problem is formulated. By analyzing the relationship between clustering and cooperation and utilizing the solutions of the knapsack problems, the above challenging optimization problem is transformed into a clustering subproblem and a content placement subproblem. To further reduce complexity, we propose an effective graph-based approach to solve the two subproblems. In the graph-based clustering approach, a node graph and a weighted graph are constructed. By setting the weights of the vertices of the weighted graph to be the incremental offloaded traffics of their corresponding complete subgraphs, the objective cluster sets can be readily obtained by using an effective greedy algorithm to search for the max-weight independent subset. In the graph-based content placement approach, a redundancy graph is constructed by removing the edges in the complete subgraphs of the node graph corresponding to the obtained cluster sets. Furthermore, we enhance the caching decisions to ensure each duplicate file is cached only once. Compared with traditional approximate solutions, our proposed graph-based approach has lower complexity. Simulation results show remarkable improvements in terms of offloaded traffic by using our proposed approach.

Collaborative quantum computation with redundant graph state

Quantum computation is a computing model based on quantum theory, which can outperform the classical computation in solving certain problems. With the increase of the complexity of quantum computing tasks, it becomes important to distribute quantum computing resources to multi-parties to cooperatively fulfill the complex tasks. Here in this paper a scheme based on the one-way quantum computing model is proposed to realize collaborative quantum computation. The standard one-way quantum computing model is based on graph states. With graph states used as resources, one can realize a universal quantum computer through using single-qubit measurements and feed-forward. In contrast to the standard one-way computation, the main resource for collaborative quantum computation is a redundant graph state (also a multi-particle highly entangled state). Unlike in the traditional graph state where each particle corresponds to a specific node, in a redundant graph state, several particles correspond to a single node, which means that each node of the graph has several redundant copies. With the help of a redundant graph state, several parties can share a graph state flexibly at will. A redundant graph state is prepared and then distributed to several parties where each of them obtains a full copy of all nodes. By communicating with each other and measuring the particles in different ways, a standard graph state is prepared and distributed among these parties. The collaborative computation then finishes through the common one-way quantum computing operations. Besides the general scheme, a concrete optical implementation of a two-party cooperative single-qubit quantum state preparation based on a six-photon redundant graph state is also put forward. Such a redundant graph state is proposed to be prepared by using the spontaneous parametric down-conversion entangled source and quantum interference. With this redundant graph state, a standard three-node graph state can be shared with the two parties in an arbitrary way. This scheme does not only make the collaborative quantum computation across several parties possible and flexible, but also guarantee the privacy of each party’s operations. This feature would be particularly useful in the case where the computing resource is obtained from an outside provider. This scheme paves the way for realizing quantum computation in more general and complicated applications.

Design Optimization Considering Guiding Template Feasibility and Redundant Via Insertion for Directed Self-Assembly

While multiple patterning lithography suffers from considerable and increasing mask manufacturing cost, next generation lithography technologies are urgently required for sub-10 nm technology nodes, where directed self-assembly (DSA) is one of the most promising candidates for contact/via layer fabrication. In addition, redundant via insertion is regarded as an important step in the circuit design flow to improve circuit reliability and yield. In this paper, we propose to adopt wire perturbation to further enhance via manufacturability and redundant via insertion rates. In addition, an improved DSA-compliant and redundant via-aware routing graph model is proposed and a systematic via graph update approach is developed to facilitate the implementation of the router. Experimental results demonstrate that compared with a state-of-the-art work, our wire perturbation approach can averagely increase inserted redundant vias by 6% and reduce unmanufacturable vias by 23% with only 0.9% wirelength overhead, and our routing graph model achieves the same performance as the one proposed in a state-of-the-art work and is much more efficient.

Scribal influence on the transmission of pre-Qin texts, as seen in excavated manuscripts

ABSTRACTThis article is directed against some Western scholars’ unfounded claim that early Chinese texts were predominantly transmitted orally (either through dictation or writing from memory), and discusses the importance of the scribe in the transmission of pre-Qin texts. Concrete evidence from excavated manuscripts suggests that early Chinese texts were primarily transmitted as copies that passed from hand to hand and text to text. The author then discusses, through handwriting analysis, examples of texts copied by a single scribe as well as multiple scribes, and how the copyists treated texts and scripts that originated from different states of the Warring States period. The article then discusses how scribes treated textual errors, omissions, and redundant graphs and texts arising during the process of textual transmission. Finally the author discusses the social status of the scribe in early China and hypothesizes that there may have already been bookshops in the Warring States period.

Distance-Aware Selective Online Query Processing Over Large Distributed Graphs

Performing online selective queries against graphs is a challenging problem due to the unbounded nature of graph queries which leads to poor computation locality. It becomes even difficult when a graph is too large to be fit in the memory. Although there have been emerging efforts on managing large graphs in a distributed and parallel setting, e.g., Pregel, HaLoop and etc, these computing frameworks are designed from the perspective of scalability instead of the query efficiency. In this work, we present our solution methodology for online selective graph queries based on the shortest path distance semantic, which finds various applications in practice. The essential intuition is to build a distance-aware index for online distance-based query processing and to eliminate redundant graph traversal as much as possible. We discuss how the solution can be applied to two types of research problems, distance join and vertex set bonding, which are distance-based graph pattern discovery and finding the structure-wise bonding of vertices, respectively.

Rigid tensegrity labelings of graphs

Tensegrity frameworks are defined on a set of points in R d and consist of bars, cables, and struts, which provide upper and/or lower bounds for the distance between their endpoints. The graph of the framework, in which edges are labeled as bars, cables, and struts, is called a tensegrity graph. It is said to be rigid in R d if it has an infinitesimally rigid realization in R d as a tensegrity framework. The characterization of rigid tensegrity graphs is not known for d ≥ 2 . A related problem is how to find a rigid labeling of a graph using no bars. Our main result is an efficient combinatorial algorithm for finding a rigid cable–strut labeling of a given graph in the case when d = 2 . The algorithm is based on a new inductive construction of redundant graphs, i.e. graphs which have a realization as a bar framework in which each bar can be deleted without increasing the degree of freedom. The labeling is constructed recursively by using labeled versions of some well-known operations on bar frameworks.

Fault Tolerant Task Re-Assignation in a Parallel-Inverters System

In this paper a task re-assignation for a Parallel-Inverters System integrated by a Master-Inverter and n Slave-Inverters is proposed such that the required power is holding if a fault occurs in sensors and actuators of the system. The considered fault scenario is a fault in the Master-Inverter or in any of the Slave-Inverters. The diagnosis capability of the system is made using the structural analysis based on graph tools. On the base of the faults signature and the redundant graphs, the Analytical Redundancy Relations are obtained for the system with two Slave-Inverters. The fault tolerant scheme which supplies the required power at the load in fault conditions is achieved changing only the references signals of the control loops for the Master-Inverter and each Slave-Inverter, and disconnecting the faulty inverter of the whole system. Thus, the addition of a reconfiguration logic with a Fault Detection and Isolation system to the conventional Parallel-Inverters System allows to satisfy the power supply at the load in fault conditions. Simulation results show the potentiality of the simple task re-assignation in the whole system.

Comment on “Efficient high-fidelity quantum computation using matter qubits and linear optics”

Recently Barrett and Kok (BK) proposed an elegant method for entangling separated matter qubits. They outlined a strategy for using their entangling operation (EO) to build graph states, the resource for one-way quantum computing. However by viewing their EO as a graph fusion event, one perceives that each successful event introduces an ideal redundant graph edge, which growth strategies should exploit. For example, if each EO succeeds with probability p=0.4 then a highly connected graph can be formed with an overhead of only about ten EO attempts per graph edge. The BK scheme then becomes competitive with the more elaborate entanglement procedures designed to permit p to approach unity.

Sensor Value Validation Based on Implicit Sensor Redundancy for Reliable Operation of Power Plants

This paper presents an innovative sensor value validation technique that can systematically explore the embedded sensor redundancies in power plants and utilize them in validating sensor values. The proposed technique uses the relationships existing among the sensor values, which is called causal relations (CRs) in this paper. A data structure called a sensor redundancy graph (SRG) is formulated to explore systematically the sensor redundancies existing throughout the plants. In the SRG, CRs are represented as nodes and two CRs having a common sensor are connected with an edge. The criteria are also provided for testing whether a given sensor relationship expressed as a CR holds with the actual measurement values. If the relationship holds, the necessary condition for the validities of the sensors participated in the CR is satisfied. A technique that can examine the compatibility of the precisions among the sensors in a CR for testing the usability of the given CR is also developed based on sensor precision indices and statistical inference. The proposed technique is applied to flow meter data provided from an actual super critical steam power plant and its application potential is demonstrated.

Using the Redundant Chords of the k -Fault-Tolerant Graph to Eliminate the Effect of Failed Components

Studies of the structural graphs of the fault-tolerant systems designed by the methods of M.F. Karavai demonstrated that if the number of the redundant graph vertices coincides with the degree k of fault-tolerance, then the number of redundant chords grows rapidly with k. Described was a method of reconfiguration of the redundant graph which is a variant of the sliding redundantization allowing one to use the redundant chords in order to eliminate the effect of some m (> k) failed chords of the graph, its value being dependent on the structure of the objective graph, the magnitude of k, and the topology of failures. Presented were examples of realizing this method which suggest that greater values of m are more frequently obtained for the longest chords of the k-fault-tolerant graph.

Effects of Graphs on Text Comprehension

Sources of information, from textbooks to multimedia encyclopedias combine text and graphics. This paper applies the construction-integration model of text comprehension to the processing of text and quantitative graphs. The model suggests that, when reading a graph that has information that is redundant with the text, a reader may directly connect the propositional information from the graph and the text due to overlap. When reading a graph that has information that elaborates on the text, a reader makes inferences to connect the propositional information from the graph and the text. Hypotheses from this model are that (1) use of an elaborative graph should produce better comprehension if the graph and text are concurrent, whereas (2) use of a redundant graph should produce better comprehension if the graph and text are separated. Three experiments tested these hypotheses by giving readers text with three types of graphs – text only, redundant graphs, and elaborative graphs — with the graphs either concurrent with (Experiment 1), before (Experiment 2A), or after (Experiment 2B) the text, then giving readers text-based or inference-based questions. In all three orders of the graph and text, performance on the inference-based questions was better in the elaborative graph condition than in the redundant graph condition. The discussion extends the application of the construction-integration model to reading text with graphs.

Partial redundancy elimination in SSA form

The SSAPRE algorithm for performing partial redundancy elimination based entirely on SSA form is presented. The algorithm is formulated based on a new conceptual framework, the factored redundancy graph, for analyzing redundancy, and representes the first sparse approach to the classical problem and on methods for its solution. With the algorithm description, theorems and their proofs are given showing that the algorithm produces the best possible code by the criteria of computational optimality and lifetime optimality of the introduced temporaries. In addition to the base algorithm, a practical implementation of SSAPRE that exhibits additional compile-time efficiencies is described. In closing, measurement statistics are provided that characterize the instances of the partial redundancy problem from a set of benchmark programs and compare optimization time spent by an implementation of SSAPRE aganist a classical partial redundancy elimination implementation. The data lend insight into the nature of partial redundancy elimination and demonstrate the expediency of this new approach.