RISE Research Radar

Computer Science Open House 2022-2025

2025

LiMNet: efficient real-time AI inference for high-frequency temporal interactions

Lodovico Giaretta, Sarunas Girdzijauskas

Summary

Lightweight Memory Network for processing high-frequency temporal interactions and capturing event causality patterns. Highly scalable: processes 1000s of interactions using single CPU core with less than 1MB memory. Applications include botnet detection, fraud/laundering detection, disease propagation tracking, fake news detection, and recommender systems.

Themes

machine-learningcybersecurity

Keywords

temporal interactions, lightweight ML, botnet detection, fraud detection

Poster

LiMNet: efficient real-time AI inference for high-frequency temporal interactions poster

Click image to open full size