# Optimizing Diffusion Rate and Label Reliability in a Graph-Based Semi-supervised Classifier

arXiv: https://arxiv.org/abs/2201.03456 · HTML: https://mlsanity.com/paper/2201.03456/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 8 of its 19 references that are in the map.
- First posted on arXiv 10 Jan 2022.

## Builds on (most cited first, 8 of 8)

- [Billion-scale similarity search with GPUs](https://mlsanity.com/paper/1702.08734/) (2017)
- [Graph Convolutional Neural Networks for Web-Scale Recommender Systems](https://mlsanity.com/paper/1806.01973/) (2018)
- [Virtual Adversarial Training: A Regularization Method for Supervised and Semi-Supervised Learning](https://mlsanity.com/paper/1704.03976/) (2017)
- [Molecular Graph Convolutions: Moving Beyond Fingerprints](https://mlsanity.com/paper/1603.00856/) (2016)
- [Trend Filtering on Graphs](https://mlsanity.com/paper/1410.7690/) (2014)
- [Analysis of label noise in graph-based semi-supervised learning](https://mlsanity.com/paper/2009.12966/) (2020)
- [Learning with Inadequate and Incorrect Supervision](https://mlsanity.com/paper/1902.07429/) (2019)
- [Identifying noisy labels with a transductive semi-supervised leave-one-out filter](https://mlsanity.com/paper/2009.11811/) (2020)

## Built on by (most cited first, 1 of 1)

- [QT-Routenet: Improved GNN generalization to larger 5G networks by fine-tuning predictions from queueing theory](https://mlsanity.com/paper/2207.06336/) (2022)

## Abstract (arXiv)

Semi-supervised learning has received attention from researchers, as it allows one to exploit the structure of unlabeled data to achieve competitive classification results with much fewer labels than supervised approaches. The Local and Global Consistency (LGC) algorithm is one of the most well-known graph-based semi-supervised (GSSL) classifiers. Notably, its solution can be written as a linear combination of the known labels. The coefficients of this linear combination depend on a parameter α, determining the decay of the reward over time when reaching labeled vertices in a random walk. In this work, we discuss how removing the self-influence of a labeled instance may be beneficial, and how it relates to leave-one-out error. Moreover, we propose to minimize this leave-one-out loss with automatic differentiation. Within this framework, we propose methods to estimate label reliability and diffusion rate. Optimizing the diffusion rate is more efficiently accomplished with a spectral representation. Results show that the label reliability approach competes with robust L1-norm methods and that removing diagonal entries reduces the risk of overfitting and leads to suitable criteria for parameter selection.