Self-Supervised Learning of Contextual Embeddings for Link Prediction in Heterogeneous Networks

Wang, Ping; Agarwal, Khushbu; Ham, Colby; Choudhury, Sutanay; Reddy, Chandan K.

doi:10.1145/3442381.3450060

Cited by 49 publications

(12 citation statements)

References 18 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Comparison methods. As mentioned in Section 2.2, the five state-of-the-art prediction methods, 𝑆𝐸𝐴𝐿 [21], 𝐺𝐴𝑇𝑁𝐸 [22], 𝐻𝑒𝐺𝐴𝑁 [23], 𝑆𝐿𝑖𝐶𝐸 [25], and 𝑃𝑀𝐸 [26] are considered as comparison methods. In addition, the missing link prediction problem can be considered as a binary classification that distinguishes between missing links and nonexistent links.…”

Section: Methodsmentioning

confidence: 99%

“…They utilized both network and node features to learn the distance measure in a coupled fashion by employing the multitask structure preserving metric learning setup. In addition, Wang et al [25] designed a self-supervised learning of contextual embedding (𝑆𝐿𝑖𝐶𝐸) model using localized attention driven mechanisms. They pre-trained their model in a self-supervised manner by introducing higher-order semantic associations and masking nodes, and then fine-tuned it for a specific link prediction task.…”

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

A Multi-Type Transferable Method for Missing Link Prediction in Heterogeneous Social Networks

Wang

Cui

Liu

et al. 2023

IEEE Trans. Knowl. Data Eng.

View full text Add to dashboard Cite

Heterogeneous social networks, which are characterized by diverse interaction types, have resulted in new challenges for missing link prediction. Most deep learning models tend to capture type-specific features to maximize the prediction performances on specific link types. However, the types of missing links are uncertain in heterogeneous social networks; this restricts the prediction performances of existing deep learning models. To address this issue, we propose a multi-type transferable method (𝑀𝑇𝑇𝑀) for missing link prediction in heterogeneous social networks, which exploits adversarial neural networks to remain robust against type differences. It comprises a generative predictor and a discriminative classifier. The generative predictor can extract link representations and predict whether the unobserved link is a missing link. To generalize well for different link types to improve the prediction performance, it attempts to deceive the discriminative classifier by learning transferable feature representations among link types. In order not to be deceived, the discriminative classifier attempts to accurately distinguish link types, which indirectly helps the generative predictor judge whether the learned feature representations are transferable among link types. Finally, the integrated 𝑀𝑇𝑇𝑀 is constructed on this minimax two-player game between the generative predictor and discriminative classifier to predict missing links based on transferable feature representations among link types. Extensive experiments show that the proposed 𝑀𝑇𝑇𝑀 can outperform state-of-the-art baselines for missing link prediction in heterogeneous social networks.

show abstract

Section: Methodsmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

A Multi-Type Transferable Method for Missing Link Prediction in Heterogeneous Social Networks

Wang

Cui

Liu

et al. 2023

IEEE Trans. Knowl. Data Eng.

View full text Add to dashboard Cite

show abstract

“…Then, the encoded node representations are further shuffled to increase the difficulty of pretext task. On heterogeneous graphs, SLiCE [96] pulls nodes closer to their closest contextual graphs, instead of explicitly contrasting nodes with the entire graph. In addition, SLiCE enriches the localized information of node embeddings via a contextual translation mechanism.…”

Section: Cross-scale Contrastmentioning

confidence: 99%

Graph Self-Supervised Learning: A Survey

Liu

Jin

Pan

et al. 2022

IEEE Trans. Knowl. Data Eng.

216

View full text Add to dashboard Cite

Deep learning on graphs has attracted significant interests recently. However, most of the works have focused on (semi-) supervised learning, resulting in shortcomings including heavy label reliance, poor generalization, and weak robustness. To address these issues, self-supervised learning (SSL), which extracts informative knowledge through well-designed pretext tasks without relying on manual labels, has become a promising and trending learning paradigm for graph data. Different from SSL on other domains like computer vision and natural language processing, SSL on graphs has an exclusive background, design ideas, and taxonomies. Under the umbrella of graph self-supervised learning, we present a timely and comprehensive review of the existing approaches which employ SSL techniques for graph data. We construct a unified framework that mathematically formalizes the paradigm of graph SSL. According to the objectives of pretext tasks, we divide these approaches into four categories: generation-based, auxiliary property-based, contrast-based, and hybrid approaches. We further describe the applications of graph SSL across various research fields and summarize the commonly used datasets, evaluation benchmark, performance comparison and open-source codes of graph SSL. Finally, we discuss the remaining challenges and potential future directions in this research field.

show abstract

“…A new time-series link prediction approach was presented which is based on irregular cellular learning automatons and evolutionary computation [41]. To outfit the link prediction problems that requires specific contextual information which may be taken out from the sub-graphs, a framework was designed by Wang et al [42] to bridge the static depiction learning approaches utilizing global information. In the current year, utilizing the relationship between phenotype and genotype, a machine learning based link prediction algorithm is also proposed [43].…”

Section: Related Workmentioning

confidence: 99%

A Novel PSO based Intelligence Link Prediction Algorithm in Real World Networks

Choudhury

Acharjee

2022

Preprint

View full text Add to dashboard Cite

Link prediction is a critical challenge to analyse social networks with applications in areas such as information retrieval, bioinformatics, e-commerce etc. To overcome the challenges of either predicting the missing links or finding the future connections, a variety of approaches have been presented so far. Similarity-based approaches, which evaluate the network structure and use the number of shared neighbours between two nodes as a key criterion to determine structural similarity , are the most efficient among these methods. In this paper, we have proposed a novel link prediction algorithm based on Particle Swarm Optimization (PSO) and implemented it on different real world datasets. Our proposed approach achieved 72.75% accuracy in Zachary karate Club, 75.26% accuracy in Bottlenose Dolphin Network, 80.35% accuracy in College Football Network, and 78.36% accuracy in Kreb’s book Network which beat the performances of traditional algorithms.

show abstract

Self-Supervised Learning of Contextual Embeddings for Link Prediction in Heterogeneous Networks

Cited by 49 publications

References 18 publications

A Multi-Type Transferable Method for Missing Link Prediction in Heterogeneous Social Networks

A Multi-Type Transferable Method for Missing Link Prediction in Heterogeneous Social Networks

Graph Self-Supervised Learning: A Survey

A Novel PSO based Intelligence Link Prediction Algorithm in Real World Networks

Contact Info

Product

Resources

About