Contribution to the Application of Machine Learning in Real-Time Scenarios (Malware Detection, Industry 4.0, and Environmental Soft-Sensors)

Stanislav Vakaruk (Université de Madrid, visiteur à MIAT)

16 déc. 2022

Résumé Machine Learning and Deep Learning methods can be applied in almost any field to solve some type of problem, usually requiring a labeled dataset (preferably large). In addition to the fact that these techniques or models learn to solve problems without explicit programming, they can solve them fast enough to be incorporated into a real-time solution. In this presentation, I will talk about a set of innovative solutions that we have found over the last three years in a variety of fields. In the field of network malware detection, we have shown that it is possible to detect cryptomining connections even if they are encrypted. In the field of Industry 4.0, we have shown that it is possible to predict the deviation of an Automatic Guided Vehicle (AGV) connected by a 5G connection to a controller on an EDGE device even under unstable network situations. In the environmental field, we have designed a Chl-a soft-sensor from low-cost variables in a freshwater concentration.;