Accurate graph classification via two-staged contrastive curriculum learning.

Given a graph dataset, how can we generate meaningful graph representations that maximize classification accuracy? Learning representative graph embeddings is important for solving various real-world graph-based tasks. Graph contrastive learning aims to learn representations of graphs by capturing the relationship between the original graph and the augmented graph. However, previous contrastive learning methods neither capture semantic information within graphs nor consider both nodes and graphs while learning graph embeddings. We propose TAG (Two-staged contrAstive curriculum learning for Graphs), a two-staged contrastive learning method for graph classification. TAG learns graph representations in two levels: node-level and graph level, by exploiting six degree-based model-agnostic augmentation algorithms. Experiments show that TAG outperforms both unsupervised and supervised methods in classification accuracy, achieving up to 4.08% points and 4.76% points higher than the second-best unsupervised and supervised methods on average, respectively.

A Novel Method for Calibration of Digital Soil Images Captured under Irregular Lighting Conditions.

Soil color is commonly used as an indicator to classify soil and identify its properties. However, color-based soil assessments are susceptible to variations in light conditions and the subjectivity of visual evaluations. This study proposes a novel method of calibrating digital images of soil, regardless of lighting conditions, to ensure accurate identification. Two different color space models, RGB and CIELAB, were assessed in terms of their potential utility in calibrating changes to soil color in digital images. The latter system was determined to be suitable, as a result of its ability to accurately reflect illuminance and color temperature. Linear regression equations relating soil color and light conditions were developed based on digital images of four different types of soil samples, each photographed under 15 different light conditions. The proposed method can be applied to calibrate variations in the soil color obtained by digital images, thus allowing for more standardized, objective, and accurate classification and evaluation of soil based on its color.