A Survey on Network Embedding

Cui, Peng; Wang, Xiao; Pei, Jian; Zhu, Wenwu

doi:10.1109/tkde.2018.2849727

Cited by 984 publications

(536 citation statements)

References 77 publications

Supporting

Mentioning

533

Contrasting

Unclassified

Order By: Relevance

“…As introduced above, ϕ is a graph encoder that takes graph as input and outputs an intermediate vector u i for any node v i . In general, the choice of ϕ is very flexible and can be instantiated by any deep representation methods as discussed in [7]. In this work, we consider a two-layer graph convolutional network GCN [16].…”

Section: Graph Encodermentioning

confidence: 99%

Learning to Hash with Graph Neural Networks for Recommender Systems

Tan

Liu

Zhao

et al. 2020

Proceedings of the Web Conference 2020

View full text Add to dashboard Cite

Recommender systems in industry generally include two stages: recall and ranking. Recall refers to efficiently identify hundreds of candidate items that user may interest in from a large volume of item corpus, while the latter aims to output a precise ranking list using complex ranking models. Recently, graph representation learning has attracted much attention in supporting high quality candidate search at scale. Despite its effectiveness in learning embedding vectors for objects in the user-item interaction network, the computational costs to infer users' preferences in continuous embedding space are tremendous. In this work, we investigate the problem of hashing with graph neural networks (GNNs) for high quality retrieval, and propose a simple yet effective discrete representation learning framework to jointly learn continuous and discrete codes. Specifically, a deep hashing with GNNs (HashGNN) is presented, which consists of two components, a GNN encoder for learning node representations, and a hash layer for encoding representations to hash codes. The whole architecture is trained end-to-end by jointly optimizing two losses, i.e., reconstruction loss from reconstructing observed links, and ranking loss from preserving the relative ordering of hash codes. A novel discrete optimization strategy based on straight through estimator (STE) with guidance is proposed. The principal idea is to avoid gradient magnification in back-propagation of STE with continuous embedding guidance, in which we begin from learning an easier network that mimic the continuous embedding and let it evolve during the training until it finally goes back to STE. Comprehensive experiments over three publicly available and one real-world Alibaba company datasets demonstrate that our model not only can achieve comparable performance compared with its continuous counterpart but also runs multiple times faster during inference.

show abstract

Section: Graph Encodermentioning

confidence: 99%

Learning to Hash with Graph Neural Networks for Recommender Systems

Tan

Liu

Zhao

et al. 2020

Proceedings of the Web Conference 2020

View full text Add to dashboard Cite

show abstract

“…While being accurate, training a supervised-learning model on the adjacency matrix (SL-A) can take some computational time and resources as the size of the molecular network increases, thus considerably differing in speed for, say, STRING-EXP (14,089 nodes and 141,629 unweighted edges) and GIANT-TN (25,689 nodes and 38,904,929 weighted edges). Worthy of note in this context is the recent excitement in deriving node embeddings for each node in a network, concisely encoding its connectivity to all other nodes, and using them as features in SL algorithms for node classification [60,[69][70][71][72][73][74] . Although we show that SL-A markedly outperforms supervised-learning on the embedding matrix (SL-E; Fig.…”

Section: Discussionmentioning

confidence: 99%

Supervised-learning is an accurate method for network-based gene classification

Liu

Mancuso

Yannakopoulos

et al. 2019

Preprint

View full text Add to dashboard Cite

Background: Assigning every human gene to specific functions, diseases, and traits is a grand challenge in modern genetics. Key to addressing this challenge are computational methods such as supervised-learning and label-propagation that can leverage molecular interaction networks to predict gene attributes. In spite of being a popular machine learning technique across fields, supervised-learning has been applied only in a few network-based studies for predicting pathway-, phenotype-, or disease-associated genes. It is unknown how supervised-learning broadly performs across different networks and diverse gene classification tasks, and how it compares to label-propagation, the widely-benchmarked canonical approach for this problem. Results:In this study, we present a comprehensive benchmarking of supervised-learning for network-based gene classification, evaluating this approach and a state-of-the-art label-propagation technique on hundreds of diverse prediction tasks and multiple networks using stringent evaluation schemes. We demonstrate that supervised-learning on a gene's full network connectivity outperforms label-propagation and achieves high prediction accuracy by efficiently capturing local network properties, rivaling label-propagation's appeal for naturally using network topology. We further show that supervised-learning on the full network is also superior to learning on node-embeddings (derived using node2vec ), an increasingly popular approach for concisely representing network connectivity. Conclusion:These results show that supervised-learning is an accurate approach for prioritizing genes associated with diverse functions, diseases, and traits and should be considered a staple of network-based gene classification workflows. The datasets and the code used to reproduce the results and add new gene classification methods have been made freely available.

show abstract

“…Comprehensive surveys of network embedding algorithms can be found elsewhere . There is an immense catalogue of algorithms, and code is distributed in a rushing pace (over 50 network embedding packages are available, many of them released during the last 2 years; https://github.com/chihming/awesome-network-embedding).…”

Section: Towards Biological Embeddingsmentioning

confidence: 99%

“…Comprehensive surveys of network embedding algorithms can be found elsewhere. 107,108,110 There is an immense catalogue of algorithms, and code is distributed in a rushing pace (over 50 network embedding packages are available, many of them released during the last 2 years; https://github.com/chihming/awesome-network-embedding). Families of successful network embedding algorithms include adjacency matrix factorizations (e.g., graph Laplacian eigenmaps), local linear embeddings, isomaps, and a series of deep learning implementations that address several scenarios, such as the case of attributed networks or the preservation of network structure and properties.…”

Section: Figurementioning

confidence: 99%

Formatting biological big data for modern machine learning in drug discovery

Duran‐Frigola

Fernández‐Torras

Bertoni

et al. 2018

WIREs Comput Mol Sci

View full text Add to dashboard Cite

Biological data is accumulating at an unprecedented rate, escalating the role of data‐driven methods in computational drug discovery. This scenario is favored by recent advances in machine learning algorithms, which are optimized for huge datasets and consistently beat the predictive performance of previous art, rapidly approaching human expert reasoning. The urge to couple biological data to cutting‐edge machine learning has spurred developments in data integration and knowledge representation, especially in the form of heterogeneous, multiplex and semantically‐rich biological networks. Today, thanks to the propitious rise in knowledge embedding techniques, these large and complex biological networks can be converted to a vector format that suits the majority of machine learning implementations. Here, we explain why this can be particularly transformative for drug discovery where, for decades, customary chemoinformatics methods have employed vector descriptors of compound structures as the standard input of their prediction tasks. A common vector format to represent biology and chemistry may push biological information into most of the existing steps of the drug discovery pipeline, boosting the accuracy of predictions and uncovering connections between small molecules and other biological entities such as targets or diseases. This article is categorized under: Computer and Information Science > Databases and Expert Systems Computer and Information Science > Chemoinformatics

show abstract

A Survey on Network Embedding

Cited by 984 publications

References 77 publications

Learning to Hash with Graph Neural Networks for Recommender Systems

Learning to Hash with Graph Neural Networks for Recommender Systems

Supervised-learning is an accurate method for network-based gene classification

Formatting biological big data for modern machine learning in drug discovery

Contact Info

Product

Resources

About