USENIX Security '23 - Diving into Robocall Content with SnorCall
Description:
Explore a groundbreaking research presentation on combating robocalls through advanced content analysis. Learn about SnorCall, an innovative framework designed to efficiently extract and classify content from large-scale robocall datasets. Discover how this tool leverages the Snorkel framework and expert-written labeling functions to achieve high-accuracy classification of call transcripts. Gain insights into the prevalence of various scam and legitimate robocall topics, referenced organizations, average scam solicitation amounts, shared infrastructure between campaigns, and the dynamics of election-related political calls. Understand the potential applications of SnorCall for regulators, carriers, anti-robocall product vendors, and researchers in developing more effective defenses against the growing robocall problem.