Abstract:Network data in real-world tends to be error-prone due to incomplete sampling, imperfect measurements, etc.; this in turn results in inaccurate results when performing network analysis or modeling, such as node classification and link prediction, on these flawed networks. In this paper, we aim to reconstruct a reliable network from a flawed, undirected, unweighted network, a process referred to network enhancement. More specifically, network enhancement aims to detect the noisy links that are observed in the network but should not exist in the real world, as well as to predict the missing links that do indeed exist in the real world yet remain unobserved. While some attempts have been made to detect either noisy links or missing links, few of these works have considered unifying these two tasks, even though they are inter-dependent and capable of mutually boosting each others’ performance. In this paper, we therefore propose E-Net, an end-to-end graph neural network model, to leverage the mutual influence of these two tasks in order to achieve both goals more effectively. On one hand, detecting noisy links can benefit the performance of missing link prediction, while on the other hand, predicting missing links can provide indirect supervision for detecting noisy link detection when the labels of these noisy links are unavailable. Moreover, by proposing a subgraph extraction mechanism based on random walk with restart, the model can be scaled up to large networks and is able to preserve the local and global structural characteristics. The experimental results on several types of large networks demonstrate that the proposed model obtains an improvement of 10.7 percent on average in terms of F1 for predicting missing links, along with an average of 3.7 percent improvement in terms of precision for detecting noisy links compared with the state-of-the-art baselines.

Enhancing the Robustness and Efficiency of Scale-free Network with Limited Link Addition.

A Topological Reconfiguration Method for Enhancing Networks Survivability with Limited Resources

A top ological reconfiguration metho d for enhancing networks survivability with limited resources

A systematic method for network topology reconfiguration with limited link additions

A near optimal solution for network topology reconfigurations with limited link resources

Enhancing network robustness for malicious attacks

Optimization of Robustness of Complex Networks

Some inverse spectral results for semi-classical Schrödinger operators

New attack strategies for complex networks

Global Disassortative Rewiring Strategy for Enhancing the Robustness of Scale-Free Networks Against Localized Attack.

Enhancing the Robustness of Scale-Free Networks: The Simulation of Cascade Failures with Adjustable Initial Load Parameters

Reconfiguring Networked Infrastructures by Adding Wireless Communication Capabilities to Selected Nodes

Optimizing the robustness of higher-low order coupled networks

Attack Vulnerability of Network Controllability.

Research on link addition strategies of interdependent networks under different coupling methods

Robustness Of Controllability For Networks Based On Edge-Attack

Robust Network Enhancement from Flawed Networks.

Efficient Topology Optimization for a Wired Networked System by Adding Wireless Communication

Adversarial Attacks to Scale-Free Networks: Testing the Robustness of Physical Criteria

Which Link Matters? Maintaining Connectivity of Uncertain Networks Under Adversarial Attack

Robustness on distributed coupling networks with multiple dependent links from finite functional components