Description:

Monday / CS Seminar
April 3rd / 12:00 PM
Mudd 3514

Title: Unleashing the Potential of Approximate Computing Systems
Speaker: Alan Zaoxing Liu

Abstract: 
Today, we are seeing an explosion in emerging compute/data-intensive applications, such as AI and data analysis services, the rollout of next-generation networks, and the growth of smart edge devices. However, the transition to a post-Moore era raises significant concerns ranging from compute and storage efficiency to carbon footprint/energy consumption. An underexplored but promising opportunity to improve the cost-performance-sustainability tradeoffs of existing computing systems is the use of approximation. Data systems may not need to calculate results with 100% precision to maintain operational reliability.
 
In this talk, I will present my research on scaling computing systems with approximation techniques for various analytical tasks across the computing stack, such as network traffic analysis and dynamic connected data processing. First, I will describe how bridging theory and practice with sketching and sampling techniques can significantly speed up network analytics under tight resource budgets. Second, I will discuss efficient algorithms and system optimizations that enable mining complex structures in large-scale graph data. The systems I have built are backed with rigorous theoretical guarantees and achieve several orders of magnitude improvements with small accuracy losses. The developed network analytics solution is the first of its kind deployed in popular open-source network processing libraries (e.g., Data Plane Development Kit), and the approximate graph systems are under evaluation in the industry (e.g., a fintech company and a cloud provider). Finally, I will chart paths to designing future approximate computing systems with heterogeneous hardware that balance performance, reliability, and sustainability.

Biography: 
Alan Zaoxing Liu is an Assistant Professor in Electrical and Computer Engineering at Boston University. His work spans computer systems, networks, and applied algorithms to co-design performant, reliable, and secure data analytics solutions across the computing stack. His recent research focuses on designing scalable and trustworthy approximate computing systems. He is a recipient of the best paper award at FAST'19 and received interdisciplinary recognitions, including ACM STOC "Best-of-Theory" plenary talk and USENIX ATC "Best-of-Rest". Previously, he did postdoctoral research at Carnegie Mellon University CyLab and received his Ph.D. in Computer Science from Johns Hopkins University.