Talos Targets Disinformation with Fake News Challenge Victory
Summary: Cisco Talos researchers, competing as "SOLAT IN THE SWEN," took first place out of more than 80 registered teams in the Fake News Challenge, whose first iteration (FNC-1) ran from December 1, 2016 to June 2, 2017 and focused solely on stance detection. Their winning submission was an ensemble using a 50/50 weighted average between gradient-boosted decision trees and a deep convolutional neural network, classifying an article body as agree, disagree, discuss, or unrelated to its headline, with code open-sourced under an Apache 2.0 license.
What's compelling: The challenge organizers framed stance detection as a step toward helping human fact checkers identify deliberate misinformation, and Talos notes its models were not perfectly accurate on their own - only the combined weighting got every test headline right. Disinformation campaigns run by organized actors to distort political sentiment are the stated threat, and the tooling that counters them is still research-grade.
Read the original at Cisco Talos ↗
- Topics
- AI Security
- Related
- SECURITY AFFAIRS AI-CYBERSECURITY NEWSLETTER ROUND 8 Cyber Risks · Nov 15, 2026Pentagon Renames CDAO ‘Super Intelligence Office’ Cyber Risks · Nov 15, 2026186 Fake AI Domains Redirect Users to Click Fix Lures, Browser Malware, and Search Hijackers Cyber Risks · Nov 12, 2026