Mask-GVAE: Blind Denoising Graphs via Partition
This work addresses the problem of blind graph denoising, which is crucial for improving the quality of connections in real-world networks like co-authorship networks, benefiting researchers and practitioners working with noisy graph data.
This paper introduces Mask-GVAE, a variational generative model designed for blind denoising of large discrete graphs. The model effectively recovers graph structures by deleting irrelevant edges and adding missing ones, outperforming competing approaches on PSNR and WL similarity across various benchmarks.
We present Mask-GVAE, a variational generative model for blind denoising large discrete graphs, in which "blind denoising" means we don't require any supervision from clean graphs. We focus on recovering graph structures via deleting irrelevant edges and adding missing edges, which has many applications in real-world scenarios, for example, enhancing the quality of connections in a co-authorship network. Mask-GVAE makes use of the robustness in low eigenvectors of graph Laplacian against random noise and decomposes the input graph into several stable clusters. It then harnesses the huge computations by decoding probabilistic smoothed subgraphs in a variational manner. On a wide variety of benchmarks, Mask-GVAE outperforms competing approaches by a significant margin on PSNR and WL similarity.