Distributed large-scale natural graph factorization

By Amr Ahmed, Nino Shervashidze, Shravan M. Narayanamurthy, Vanja Josifovski, Alex Smola. 2013

706 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Computer Networks and Communications, Statistical and Nonlinear Physics, Graph Learning

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Computer Networks and Communications

In Statistical and Nonlinear Physics

In Graph Learning

Citations in the corpus

LINE: Large-scale Information Network Embedding (2015, methodology): “Among the most recent literature is a technique called graph factorization [1].”
GADMM: Fast and Communication Efficient Framework for Distributed Machine Learning (2019, background): “Distributed optimization plays a pivotal role in distributed machine learning applications (Ahmed et al., 2013; Dean et al., 2012; Li et al., 2013, 2014) that commonly aims to minimize 1 N =1 f n ( θ ) with N workers.”
Representation Learning on Graphs: Methods and Applications (2017, methodology): “For example, the decoder might predict the existence of edges between nodes, given their embeddings [1, 35], or it might predict the community that a node belongs to in the graph [28, 34] (Figure 3).”
FastGCN: Fast Learning with Graph Convolutional Networks via Importance Sampling (2018, background): “A major category consists of factorization based algorithms that yield the embedding through matrix factorizations; see, e.g., Roweis & Saul (2000); Belkin & Niyogi (2001); Ahmed et al. (2013); Cao et al. (2015); Ou et al. (2016).”
A Comprehensive Survey of Graph Embedding: Problems, Techniques, and Applications (2017, methodology): “A(1= P j (A ij+ A ji))A GraRep [20] Wk ij = log( A^k ij P t A^k ij ) log( jV ), where A^ = D 1S, D ij= ˆ P p A ip;i= j 0;i6= j Eq. 5 CMF [42] PPMI Eq. 5 TADW [56] PMI Eq. 5 with text feature matrix [23] A y = argmin P eij2E (A ij <y i;y j>) 2 + 2 P i ky ik 2 MMDW [48] PMI Eq. 5 + O(SVM class”
GraRep: Learning Graph Representations with Global Structural Information (2015, methodology): “Ahmed et al. [1] proposed a graph factorization method, which used stochastic gradient descent to optimize matrices from large graphs.”
Parameter Server for Distributed Machine Learning (2013, methodology): “Of course, beyond these general frameworks, numerous systems have been developed that target specific applications, such as [3, 13, 24, 22, 28, 10, 15].”
DynGEM: Deep Embedding Method for Dynamic Graphs (2018, methodology): “…demonstrated that graph embedding is superior to alternatives in many supervised learning tasks, such as node classiﬁcation, link prediction and graph reconstruction [Ahmed et al. , 2013; Perozzi et al. , 2014; Cao et al. , 2015; Tang et al. , 2015; Grover and Leskovec, 2016; Ou et al. , 2016].”
On the Equivalence between Positional Node Embeddings and Structural Graph Representations (2019, background): “, matrix factorization (Belkin & Niyogi, 2002; Cao et al., 2015; Ahmed et al., 2013; Ou et al., 2016), implicit matrix factorization (Mikolov et al.”
Graph Learning: A Survey (2021, methodology): “There are some other approaches such as regularized Gaussian matrix factorization [83], lowrank matrix factorization [84], for solving SVD.”
Heterogeneous Graph Neural Networks for Malicious Account Detection (2018, methodology): “For example, some methods directly use the adjacency matrix [1, 3], t -th order adjacency matrix [6], and others simulate random walks by approximating the high order adjacency matrix in a randomized manner [12, 18].”
Graph embedding on biomedical networks: methods, applications and evaluations (2019, methodology): “(Ahmed et al., 2013) Yang et al. (2014) and”
Exploiting Edge Features in Graph Neural Networks. (2018, methodology): “Another class of approaches focus on employing a ﬂexible, stochastic measure of node similarity based on random walks, e.g. DeepWalk [22], node2vec [2], LINE [26], HARP [9], etc.”
Graph Representation Learning (2020, methodology): “1Of course, Ahmed et al. [Ahmed et al., 2013] were not the first researchers to propose factorizing an adjacency matrix, but they were the first to present a scalable O(|E”
dyngraph2vec: Capturing Network Dynamics using Dynamic Graph Representation Learning (2018, background): “Similarly, Graph Factorization [10] approximates the edge weight with the dot product of the nodes’ embeddings, thus preserving distance in the inner product space.”
Embedding Temporal Network via Neighborhood Formation (2018, methodology): “, by constructing and embedding the affinity graph into a low dimensional space [3, 13, 21, 23], or by applying matrix factorization to find the low-dimensional embedding [1].”
Label Informed Attributed Network Embedding (2017, background): “Ahmed et al. [2] proposed an asynchronous distributed matrix factorization algorithm for large-scale graphs.”