CICS Faculty Marco Serafini and Arjun Guha Receive Facebook Systems for Machine Learning Research Award
University of Massachusetts Amherst College of Information and Computer Sciences (CICS) Assistant Professor Marco Serafini and Associate Professor Arjun Guha recently received Facebook's Systems for Machine Learning research award for their proposal, "Massively Parallel Graph Sampling on GPUs."
The proposal describes how advanced data science applications, from social networks to knowledge bases and data integration, manage and analyze complex connected data by representing it as a graph. Although graph neural networks (GNNs) have become a popular technique to extract insights from graphs, the researchers point out that current methods used to train the models are still inefficient. To address these issues, the team will focus on developing a more efficient graph sampling system that relies on graphics processing units (GPUs) while abstracting away the complexity of their low-level behavior to machine learning experts.
Their project will be executed over one year and will build on the expertise drawn from developing Arabesque, a system that introduced the first high-level API and runtime for distributed graph mining. "Graph neural networks are becoming essential for applications like recommendations and knowledge bases, since they allow programmers to leverage the structure present in the data," says Serafini. "But, system support for GNNs is still in its infancy. Our project will develop a new execution runtime that will enable extracting training data from graphs using massively-parallel GPUs instead of regular CPUs." Serafini and Guha expect to later develop a domain-specific language to support a variety of GNN algorithms.
Facebook launched the Systems for Machine Learning proposal program with the goal of funding impactful solutions in the areas of developer toolkits, compilers/code generation, system architecture, memory technologies, and machine learning accelerator support. This year, Serafini and Guja, among ten other teams, were selected from over 167 proposals submitted from more than 100 universities. All winners are invited to the next AI Systems Faculty Summit in Fall 2020.
Marco Serafini currently works in the UMass DREAM Lab (Data systems Research for Exploration, Analytics, and Modeling), where his research lies at the intersection of database systems, distributed systems and data science. Prior to joining the CICS faculty, he was a senior scientist at the Qatar Computing Research Institute and a postdoctoral research fellow at Yahoo! Research in Barcelona. He received a PhD in computer science from the Technical University of Darmstadt, Germany, where his thesis was nominated for best thesis awards by the German, Swiss, and Austrian computer science societies and the German chapter of the ACM.
Arjun Guha, through the tools and principles of programming languages, focuses on making software more secure and more reliable in a variety of domains, including systems, networking, robotics and web programming. Along with other CICS faculty, he leads the PLASMA (Programming Languages and Systems at Massachusetts) lab, which works in a variety of areas spanning programming languages, software engineering, and systems. He received a PhD in computer science from Brown University in 2012 and a BA in computer science from Grinnell College in 2006.