Arya - Arbitrary Graph Pattern Mining with Decomposition-based Sampling

Explore an innovative approach to graph pattern mining in this 15-minute conference talk from NSDI '23. Discover Arya, an ultra-fast approximate graph pattern miner that combines novel graph decomposition theory with edge sampling-based approximation. Learn how this groundbreaking system can detect and count arbitrary patterns in graphs with up to tens of billions of edges, a scale previously only achievable on supercomputers. Understand the Error-Latency Profile (ELP) feature that allows users to configure mining task durations based on different error targets. Examine Arya's impressive performance, outpacing existing exact and approximate pattern mining solutions by up to five orders of magnitude. Gain insights into its capability to support graphs with 5 billion edges on a single machine and scale to 10-billion-edge graphs on a 32-server testbed. Presented by researchers from Boston University and the University of Wisconsin-Madison, this talk offers valuable knowledge for those interested in advanced graph processing techniques and big data analytics.