This project is aimed towards building a framework which monitors incoming ATLAS Computing Operations data for anomalies and then autonomously, acts on this information by either solving the problem or by proposing the best solution. The solution proposed for this has at its heart two machine learning models-

  • One a recurrent network to actually predict anomalies in the incoming real-time data. These predicted anomalies will then be analysed to come up with a list of potential solutions to the problem
  • Second a reinforcement learning algorithm to rank the proposed solutions based on a feedback. This algorithm will provide increasingly efficient outputs over time, thus improving the overall efficiency of the framework.

Organization

Student

Vyom Sharma

Mentors

  • Alessandro Di Girolamo
  • Mario Lassnig
close

2017