Search

We propose AGS-GNN, a novel attribute-guided sampling algorithm for Graph Neural Networks (GNNs) that exploits node features and connectivity structure of a graph while simultaneously adapting for both homophily and heterophily in graphs. (In homophilic graphs vertices of the same class are more likely to be connected, and vertices of different classes tend to be linked in heterophilic graphs.) While GNNs have been successfully applied to homophilic graphs, their application to heterophilic graphs remains challenging. The best-performing GNNs for heterophilic graphs do not fit the sampling paradigm, suffer high computational costs, and are not inductive. We employ samplers based on feature-similarity and feature-diversity to select subsets of neighbors for a node, and adaptively capture information from homophilic and heterophilic neighborhoods using dual channels. Currently, AGS-GNN is the only algorithm that we know of that explicitly controls homophily in the sampled subgraph through similar and diverse neighborhood samples. For diverse neighborhood sampling, we employ submodularity, which was not used in this context prior to our work. The sampling distribution is pre-computed and highly parallel, achieving the desired scalability. Using an extensive dataset consisting of 35 small ($\le$ 100K nodes) and large (>100K nodes) homophilic and heterophilic graphs, we demonstrate the superiority of AGS-GNN compare to the current approaches in the literature. AGS-GNN achieves comparable test accuracy to the best-performing heterophilic GNNs, even outperforming methods using the entire graph for node classification. AGS-GNN also converges faster compared to methods that sample neighborhoods randomly, and can be incorporated into existing GNN models that employ node or graph sampling., Comment: The paper has been accepted to KDD'24 in the research track

We describe a parallel approximation algorithm for maximizing monotone submodular functions subject to hereditary constraints on distributed memory multiprocessors. Our work is motivated by the need to solve submodular optimization problems on massive data sets, for practical applications in areas such as data summarization, machine learning, and graph sparsification. Our work builds on the randomized distributed RandGreedI algorithm, proposed by Barbosa, Ene, Nguyen, and Ward (2015). This algorithm computes a distributed solution by randomly partitioning the data among all the processors and then employing a single accumulation step in which all processors send their partial solutions to one processor. However, for large problems, the accumulation step could exceed the memory available on a processor, and the processor which performs the accumulation could become a computational bottleneck. Here, we propose a generalization of the RandGreedI algorithm that employs multiple accumulation steps to reduce the memory required. We analyze the approximation ratio and the time complexity of the algorithm (in the BSP model). We also evaluate the new GreedyML algorithm on three classes of problems, and report results from massive data sets with millions of elements. The results show that the GreedyML algorithm can solve problems where the sequential Greedy and distributed RandGreedI algorithms fail due to memory constraints. For certain computationally intensive problems, the GreedyML algorithm can be faster than the RandGreedI algorithm. The observed approximation quality of the solutions computed by the GreedyML algorithm closely matches those obtained by the RandGreedI algorithm on these problems., Comment: We have learned that we incorrectly applied Lemma 4.1 in Lemma 4.2. Thus, the proof of Lemma 4.2 is incorrect, and the overall approximation guarantee in Theorem 4.4 we claimed is not right

We design and implement two single-pass semi-streaming algorithms for the maximum weight $k$-disjoint matching ($k$-DM) problem. Given an integer $k$, the $k$-DM problem is to find $k$ pairwise edge-disjoint matchings such that the sum of the weights of the matchings is maximized. For $k \geq 2$, this problem is NP-hard. Our first algorithm is based on the primal-dual framework of a linear programming relaxation of the problem and is $\frac{1}{3+\varepsilon}$-approximate. We also develop an approximation preserving reduction from $k$-DM to the maximum weight $b$-matching problem. Leveraging this reduction and an existing semi-streaming $b$-matching algorithm, we design a $(\frac{1}{2+\varepsilon})(1 - \frac{1}{k+1})$-approximate semi-streaming algorithm for $k$-DM. For any constant $\varepsilon > 0$, both of these algorithms require $O(nk \log_{1+\varepsilon}^2 n)$ bits of space. To the best of our knowledge, this is the first study of semi-streaming algorithms for the $k$-DM problem. We compare our two algorithms to state-of-the-art offline algorithms on 95 real-world and synthetic test problems, including thirteen graphs generated from data center network traces. On these instances, our streaming algorithms used significantly less memory (ranging from 6$\times$ to 512$\times$ less) and were faster in runtime than the offline algorithms. Our solutions were often within 5% of the best weights from the offline algorithms. We highlight that the existing offline algorithms run out of 1 TB memory for most of the large instances ($>1$ billion edges), whereas our streaming algorithms can solve these problems using only 100 GB memory for $k=8$., Comment: 24 pages, To appear in ESA 2024

We describe a generative network model of the architecture of the enteric nervous system (ENS) in the colon employing data from images of human and mouse tissue samples obtained through confocal microscopy. Our models combine spatial point pattern analysis with graph generation to characterize the spatial and topological properties of the ganglia (clusters of neurons and glial cells), the inter-ganglionic connections, and the neuronal organization within the ganglia. We employ a hybrid hardcore-Strauss process for spatial patterns and a planar random graph generation for constructing the spatially embedded network. We show that our generative model may be helpful in both basic and translational studies, and it is sufficiently expressive to model the ENS architecture of individuals who vary in age and health status. Increased understanding of the ENS connectome will enable the use of neuromodulation strategies in treatment and clarify anatomic diagnostic criteria for people with bowel motility disorders., Comment: 17 pages, 5 figures

A thorough understanding of the neuroanatomy of peripheral nerves is required for a better insight into their function and the development of neuromodulation tools and strategies. In biophysical modeling, it is commonly assumed that the complex spatial arrangement of myelinated and unmyelinated axons in peripheral nerves is random, however, in reality the axonal organization is inhomogeneous and anisotropic. Present quantitative neuroanatomy methods analyze peripheral nerves in terms of the number of axons and the morphometric characteristics of the axons, such as area and diameter. In this study, we employed spatial statistics and point process models to describe the spatial arrangement of axons and Sinkhorn distances to compute the similarities between these arrangements (in terms of first- and second-order statistics) in various vagus and pelvic nerve cross-sections. We utilized high-resolution TEM images that have been segmented using a custom-built high-throughput deep learning system based on a highly modified U-Net architecture. Our findings show a novel and innovative approach to quantifying similarities between spatial point patterns using metrics derived from the solution to the optimal transport problem. We also present a generalizable pipeline for quantitative analysis of peripheral nerve architecture. Our data demonstrate differences between male- and female-originating samples and similarities between the pelvic and abdominal vagus nerves., Comment: 10 figures

Recent advances in quantum hardware offer new approaches to solve various optimization problems that can be computationally expensive when classical algorithms are employed. We propose a hybrid quantum-classical algorithm to solve a dynamic asset allocation problem where a target return and a target risk metric (expected shortfall) are specified. We propose an iterative algorithm that treats the target return as a constraint in a Markowitz portfolio optimization model, and dynamically adjusts the target return to satisfy the targeted expected shortfall. The Markowitz optimization is formulated as a Quadratic Unconstrained Binary Optimization (QUBO) problem. The use of the expected shortfall risk metric enables the modeling of extreme market events. We compare the results from D-Wave's 2000Q and Advantage quantum annealers using real-world financial data. Both quantum annealers are able to generate portfolios with more than 80% of the return of the classical optimal solutions, while satisfying the expected shortfall. We observe that experiments on assets with higher correlations tend to perform better, which may help to design practical quantum applications in the near term.

We design new serial and parallel approximation algorithms for computing a maximum weight $b$-matching in an edge-weighted graph with a submodular objective function. This problem is NP-hard; the new algorithms have approximation ratio $1/3$, and are relaxations of the Greedy algorithm that rely only on local information in the graph, making them parallelizable. We have designed and implemented Local Lazy Greedy algorithms for both serial and parallel computers. We have applied the approximate submodular $b$-matching algorithm to assign tasks to processors in the computation of Fock matrices in quantum chemistry on parallel computers. The assignment seeks to reduce the run time by balancing the computational load on the processors and bounding the number of messages that each processor sends. We show that the new assignment of tasks to processors provides a four fold speedup over the currently used assignment in the NWChemEx software on $8000$ processors on the Summit supercomputer at Oak Ridge National Lab., Comment: 10 pages, accepted for SIAM ACDA 21

Weaknesses in computer systems such as faults, bugs and errors in the architecture, design or implementation of software provide vulnerabilities that can be exploited by attackers to compromise the security of a system. Common Weakness Enumerations (CWE) are a hierarchically designed dictionary of software weaknesses that provide a means to understand software flaws, potential impact of their exploitation, and means to mitigate these flaws. Common Vulnerabilities and Exposures (CVE) are brief low-level descriptions that uniquely identify vulnerabilities in a specific product or protocol. Classifying or mapping of CVEs to CWEs provides a means to understand the impact and mitigate the vulnerabilities. Since manual mapping of CVEs is not a viable option, automated approaches are desirable but challenging. We present a novel Transformer-based learning framework (V2W-BERT) in this paper. By using ideas from natural language processing, link prediction and transfer learning, our method outperforms previous approaches not only for CWE instances with abundant data to train, but also rare CWE classes with little or no data to train. Our approach also shows significant improvements in using historical data to predict links for future instances of CVEs, and therefore, provides a viable approach for practical applications. Using data from MITRE and National Vulnerability Database, we achieve up to 97% prediction accuracy for randomly partitioned data and up to 94% prediction accuracy in temporally partitioned data. We believe that our work will influence the design of better methods and training models, as well as applications to solve increasingly harder problems in cybersecurity., Comment: Under submission to KDD 2021 Applied Data Science Track

We consider the maximum vertex-weighted matching problem (MVM) for non-bipartite graphs. In earlier work we have described a 2/3-approximation algorithm for the MVM on bipartite graphs (Dobrian, Halappanavar, Pothen and Al-Herz, SIAM J. Scientific Computing, 2019). Here we show that a 2/3-approximation algorithm for MVM on non-bipartite graphs can be obtained by restricting the length of augmenting paths to at most three. The algorithm has time complexity $O(m \log \Delta + n \log n)$, where $n$ is the number of vertices, $m$ is the number of edges, and $\Delta$ is the maximum degree of a vertex. The approximation ratio of the algorithm is obtained by considering failed vertices, i.e., vertices that the approximation algorithm fails to match but the exact algorithm does. We show that there are two distinct heavier matched vertices that we can charge each failed vertex to. Our proof techniques characterize the structure of augmenting paths in a novel way. We have implemented the 2/3-approximation algorithm and show that it runs in under a minute on graphs with tens of millions of vertices and hundreds of millions of edges. We compare its performance with five other algorithms: an exact algorithm for MVM, an exact algorithm for the maximum edge-weighted matching (MEM) problem, as well as three approximation algorithms. In our test set of nineteen problems, there are graphs on which the exact algorithms fail to terminate in 100 hours. The new 2/3-approximation algorithm for MVM outperforms the other approximation algorithms by either being faster (often by orders of magnitude) or obtaining better weights.

We present the AMPS algorithm, a finite element solution method that combines principal submatrix updates and Schur complement techniques, well-suited for interactive simulations of deformation and cutting of finite element meshes. Our approach features real-time solutions to the updated stiffness matrix systems to account for interactive changes in mesh connectivity and boundary conditions. Updates are accomplished by an augmented matrix formulation of the stiffness equations to maintain its consistency with changes to the underlying model without refactorization at each timestep. As changes accumulate over multiple simulation timesteps, the augmented solution algorithm enables tens or hundreds of updates per second. Acceleration schemes that exploit sparsity, memoization and parallelization lead to the updates being computed in real-time. The complexity analysis and experimental results for this method demonstrate that it scales linearly with the problem size. Results for cutting and deformation of 3D elastic models are reported for meshes with node counts up to 50,000, and involve models of astigmatism surgery and the brain., Comment: 20 pages, 9 figures, 3 tables

We consider the maximum vertex-weighted matching problem (MVM), in which non-negative weights are assigned to the vertices of a graph, the weight of a matching is the sum of the weights of the matched vertices, and we are required to compute a matching of maximum weight. We describe an exact algorithm for MVM with $O(|V|\, |E|)$ time complexity, and then we design a $2/3$-approximation algorithm for MVM on bipartite graphs by restricting the length of augmenting paths to at most three. The latter algorithm has time complexity $O(|E| + |V| \log |V|)$. The approximation algorithm solves two MVM problems on bipartite graphs, each with weights only on one vertex part, and then finds a matching from these two matchings using the Mendelsohn-Dulmage Theorem. The approximation ratio of the algorithm is obtained by considering failed vertices, i.e., vertices that the approximation algorithm fails to match but the exact algorithm does. We show that at every step of the algorithm there are two distinct heavier vertices that we can charge each failed vertex to. We have implemented the $2/3$-approximation algorithm for MVM and compare it with four other algorithms: an exact MEM algorithm, the exact MVM algorithm, a $1/2$-approximation algorithm for MVM, and a scaling-based $(1-\epsilon)$-approximation algorithm for MEM. We also show that MVM problems should not be first transformed to MEM problems and solved using exact algorithms for the latter, since this transformation can increase runtimes by several orders of magnitude.

We present AMPS, an augmented matrix approach to update the solution to a linear system of equations when the matrix is modified by a few elements within a principal submatrix. This problem arises in the dynamic security analysis of a power grid, where operators need to perform N - k contingency analysis, i.e., determine the state of the system when exactly k links from N fail. Our algorithms augment the matrix to account for the changes in it, and then compute the solution to the augmented system without refactoring the modified matrix. We provide two algorithms, a direct method, and a hybrid direct-iterative method for solving the augmented system. We also exploit the sparsity of the matrices and vectors to accelerate the overall computation. We analyze the time complexity of both algorithms, and show that it is bounded by the number of nonzeros in a subset of the columns of the Cholesky factor that are selected by the nonzeros in the sparse right-hand-side vector. Our algorithms are compared on three power grids with PARDISO, a parallel direct solver, and CHOLMOD, a direct solver with the ability to modify the Cholesky factors of the matrix. We show that our augmented algorithms outperform PARDISO (by two orders of magnitude), and CHOLMOD (by a factor of up to 5). Further, our algorithms scale better than CHOLMOD as the number of elements updated increases. The solutions are computed with high accuracy. Our algorithms are capable of computing N - k contingency analysis on a 778 thousand bus grid, updating a solution with k = 20 elements in 16 milliseconds on an Intel Xeon processor., Comment: 19 pages, 4 figures, 2 tables, SIAM Journal on Scientific Computing