Attila Mitcsenkov

Attila Mitcsenkov
About

Researcher and engineer with 10+ years experience in the field of telecommunications. I have a strong theoretical background in optimization, operations research and machine learning, holding a PhD for applying these in various network optimization & analysis.

Currently working for Ericsson as telecommunications domain expert & data scientist / analyst in the fields of VoLTE Customer Experience & Network Monitoring, being the system engineer/architect of a team developing a Machine Learning based anomaly detection application. https://www.ericsson.com/en/digital-services/offerings/network-automation/expert-analytics

I was involved in various research projects, with fixed and mobile access networks, including network planning, design, dimensioning and techno-economics, and Work Package leader of European research projects FP7 COMBO.Long-term successful cooperation with Magyar Telekom, contributing as expert to R&D projects on core and transport network design. 


Talk
Finding the needle in the haystack: transform reactive analytics to proactive monitoring

Level: General

Ericsson has a number of network analytics product, among which Ericsson Expert Analytics, due to its depth of processing is a strong contender in generating data volume; which makes it a “powerful” tool as it helps You to develop a good understanding of virtually anything in the network that You would like to analyze.However, making a powerful tool also useful needs it to be able to highlight those issues that You were not even aware of. Anomaly detection capabilities will become fundamental building blocks of telecommunication network monitoring and analytics software: automation is built on the “discover, understand, resolve” exercise, while “discovery” is anomaly detection itself.The journey from reactive analytics to proactive monitoring was really a transformation and a mindset change; with many lessons learned on how to introduce AI/ML into such a large and highly respected company’s products. Lessons on how data science (how to detect what really matters); lessons on adding new capabilities to the software and architecture; make the customers understand what they get and how is it more intelligent than it was.