GRB(Graph Robustness Benchmark)

图鲁棒性基准 (GRB) 对图机器学习模型的对抗鲁棒性提供可扩展、统一、模块化和可重复的评估。 GRB 精心设计了数据集、统一的评估管道、模块化编码框架和可重复的排行榜，它们促进了图对抗学习的发展，总结了现有进展并为未来的研究提供了见解。 GitHub：https://github.com/thudm/grb

The Graph Robustness Benchmark (GRB) provides scalable, unified, modular, and reproducible evaluations for the adversarial robustness of graph machine learning models. GRB has meticulously designed datasets, a unified evaluation pipeline, a modular coding framework, and reproducible leaderboards, which promote the development of graph adversarial learning, summarize existing research advancements, and provide insights for future studies. GitHub: https://github.com/thudm/grb

搜集汇总

背景与挑战

背景概述

GRB是一个图鲁棒性基准，用于评估图机器学习模型的对抗鲁棒性，提供可扩展、统一和模块化的评估框架。它由清华大学、浙江大学、复旦大学和Microsoft Research于2021年发布，旨在推动图对抗学习研究并总结进展。