NESTOR DC6: Introduction

NESTOR DC:6 Introduction Imran Chowdhury Dipto LinkedIn: https://www.linkedin.com/in/imran-chowdhury-dipto1bab5b203/ PhD in Electrical, Electronics and Communications Engineering December 10, 2024 About Me •Education •MSc Data Science with Placement, Manchester Metropolitan University, UK •BSc IT, University of Derby, UK •Work Experiences •Research Assistant at Manchester Metropolitan University •Data Services Analyst at Manchester University NHS Foundation Trust •Library Experience Assistant at Manchester Metropolitan University •Research Interests •AI, Computer Vision, Machine Learning and Deep Learning •Hobbies •Playing Video Games, football fan, travelling PhD Project •Title: AI-based power saving and distributed sensing/failure prevention from optical telemetry •Host Institution and Enrolment Status: Politecnico di Torino, Enrolled for the 1st Year from 14th November 2024 •Supervisor: Professor Vittorio Curri •WP:WP2–Digital Twin and Real Time control of Physical Layer •Planned Secondments •SMO, Raffaele Corsini, 8 Months •BT, Asif Iqbal, 2 months Technical Discussion •Current Activities •Shadowing the existing works of my senior Colleagues •Completing training courses on Hard Skills and Soft Skills •Hard Skills: Programming, Research Design, Mathematics etc •Soft Skills: Communication, Time Management etc •Will complete a total of 140 hours of Training •100 Hours of Hard Skills •40 Hours of Soft Skills What have I learned so far •Reviewed Work •Title: AI-aided Tomography Project •Use Case 1: Network Elements: Identification and localization of Performance degradation •Task: Fault detections in Amplifiers •Use Case 2: Fiber Cable Issues, identification and localization •Task: Multi event detection •Results: ML models showing decent predictive capabilities Objectives •Develop Methods for extracting useful performance monitoring data •Classify phenomenon like Earthquake, traffic etc using AI •Use a smart grid approach to integrate geolocation data to the classifications obtained •Implement Machine Learning methods to aid the optical control for minimizing power consumption Expected Results •Extraction of quality data •Classification of sensed phenomenon using DNN techniques •Geographical localization using various Light Paths •Energy-Efficient optical controller using ML Learning Plan •Gain domain knowledge related to optical telemetry •Learn about Time Series Data and ML models needed to handle them •Improve both technical and soft skills by attending training and secondments •Identify and participate in additional training and secondment opportunities Thank You