핵심 개념
AutoGL is the first dedicated library for automated machine learning on graphs, providing a comprehensive pipeline for graph tasks.
초록
AutoGL addresses the need for automated machine learning on graphs.
The library consists of a three-layer architecture with a fully automated pipeline.
Functional modules include auto feature engineering, neural architecture search, hyper-parameter optimization, model training, and auto ensemble.
AutoGL supports various graph tasks such as node classification, link prediction, graph classification, heterogeneous node classification, etc.
AutoGL-light is a more lightweight version focusing on core functionalities and flexibility.
NAS-Bench-Graph is integrated to provide a benchmark for graph NAS methods.
통계
최근 몇 년 동안 그래프에 대한 기계 학습에 대한 연구 관심이 증가했습니다.
AutoGL은 그래프에 대한 자동화된 기계 학습 라이브러리입니다.
AutoGL은 PyTorch Geometric 및 Deep Graph Library와 호환됩니다.
인용구
"AutoGL is the first dedicated library for automated machine learning on graphs."
"The library consists of a three-layer architecture with a fully automated pipeline."