GRAPE for fast and scalable graph processing and random-walk-based embedding

Abstract Graph representation learning methods opened new avenues for addressing complex, real-world problems represented by graphs. However, many graphs used in these applications comprise millions of nodes and billions edges are beyond the capabilities current software implementations. We present GRAPE (Graph Representation Learning, Prediction Evaluation), a resource graph processing embedding that is able to scale with big using specialized smart data structures, algorithms, fast parallel implementation random-walk-based methods. Compared state-of-the-art resources, shows an improvement orders magnitude empirical space time complexity, as well competitive edge- node-label prediction performance. comprises approximately 1.7 million well-documented lines Python Rust code provides 69 node-embedding methods, 25 inference models, collection efficient graph-processing utilities, over 80,000 from literature other sources. Standardized interfaces allow seamless integration third-party libraries, while ready-to-use modular pipelines permit easy-to-use evaluation graph-representation-learning therefore also positioning performs fair comparison between libraries embedding.