Abstract: Billions and billions of events are happening on Twitter every day. It is a big challenge to detect spam from these events in real time. Linyun Fu offers an overview of Bubbler, a real time event aggregation and ranking system built on top of the Heron stream computing framework. Bubbler surfaces suspicious spikes of activities to help fight spam on the platform.

Linyun describes Bubbler from the perspective of it being a solution to the general stream aggregation and ranking problem, and focuses on the way to overcome the technical difficulties in unique value counting and large scale ranking.

Bio: Linyun Fu is a software engineer working at Twitter. Linyun holds a PhD degree from Rensselaer Polytechnic Institute and a master's and a bachelor's from Shanghai Jiao Tong University. Linyun's area of interest is large scale anti-spam systems and tools.

Date: Monday September 10, 2018
Time: 3:30pm
Rooms: Engineering Physics 204/202, CDA Harbor Center 240B, IFCHE 301