Abstract:Graph-based semi-supervised learning, particularly in the context of extremely sparse labeled data, often suffers from degenerate solutions where label functions tend to be nearly constant across unlabeled data. In this paper, we introduce Variance-enlarged Poisson Learning (VPL), a simple yet powerful framework tailored to alleviate the issues arising from the presence of degenerate solutions. VPL incorporates a variance-enlarged regularization term, which induces a Poisson equation specifically for unlabeled data. This intuitive approach increases the dispersion of labels from their average mean, effectively reducing the likelihood of degenerate solutions characterized by nearly constant label functions. We subsequently introduce two streamlined algorithms, V-Laplace and V-Poisson, each intricately designed to enhance Laplace and Poisson learning, respectively. Furthermore, we broaden the scope of VPL to encompass graph neural networks, introducing Variance-enlarged Graph Poisson Networks (V-GPN) to facilitate improved label propagation. To achieve a deeper understanding of VPL's behavior, we conduct a comprehensive theoretical exploration in both discrete and variational cases. Our findings elucidate that VPL inherently amplifies the importance of connections within the same class while concurrently tempering those between different classes. We support our claims with extensive experiments, demonstrating the effectiveness of VPL and showcasing its superiority over existing methods. The code is available at https://github.com/hitcszx/VPL.

Semi-supervised Learning on Large Graphs: is Poisson Learning a Game-Changer?

Variance-enlarged Poisson Learning for Graph-based Semi-Supervised Learning with Extremely Sparse Labeled Data

Convergence rates for Poisson learning to a Poisson equation with measure data

Improved Graph-based semi-supervised learning Schemes

Semi-supervised Learning on Graphs with Generative Adversarial Nets

Consistency of semi-supervised learning, stochastic tug-of-war games, and the p-Laplacian

Large-Scale Graph-Based Semi-Supervised Learning Via Tree Laplacian Solver

Large-Scale Robust Semisupervised Classification

Towards Robust Graph Neural Networks against Label Noise

Efficient And Robust Semi-Supervised Learning Over A Sparse-Regularized Graph

Semi-Supervised Laplace Learning on Stiefel Manifolds

A nonlinear diffusion method for semi-supervised learning on hypergraphs

A General Graph-Based Semi-Supervised Learning with Novel Class Discovery.

Graph-Based Semi-Supervised Segregated Lipschitz Learning

Hypergraph based semi-supervised learning algorithms applied to speech recognition problem: a novel approach

Probabilistic Semi-Supervised Learning via Sparse Graph Structure Learning

Distributionally Robust Semi-Supervised Learning Over Graphs

Analysis of $p$-Laplacian Regularization in Semi-Supervised Learning

Deeper Insights Into Graph Convolutional Networks for Semi-Supervised Learning

LPNL: Scalable Link Prediction with Large Language Models

A Poincaré Inequality and Consistency Results for Signal Sampling on Large Graphs