The agenda for the seminar this week is listed below:
Title: Data Through the Window
Abstract:
Dynamic Graph analysis has become a valuable technique to analyze the content from Online Social Networks (OSN). The ever-changing of OSN content requires new and efficient techniques and mechanism to process the OSN content to understand not only the behaviors of OSN users but also the evolution of OSN. In this context, windowing is a key mechanism that allows analyzing dynamically the OSN evolution and detects immediately changes when they are produced. However, is not an easy task define and apply the right windowing mechanism because of the OSN content are generated in different places at different rates, the content represents not only relationships user to user but also relationships user to content and the content generate several reactions among OSN members. In our past work, we used this technique of time windows to analyze a dataset composed by logs of alerts, failures and events from a small closed network, and then create a machine learning model, based on that analysis, capable to predict failures in real-time on that network. During our stay, we are driving research to define a windowing mechanism and some metrics that allow us to characterize an OSN dataset and study its evolution.
Bio:
Hugo A. Parada G. (hugoalexer.parada@upm.es) earned his Ph.D. in telecommunications engineering at UPM in 2010. During the period of 2011 –2013, he had taken part in a fellowship program at Carlos III University where he had developed his work on learning analytics area. In 2013 – 2015, Dr. Parada was part of the UPM research team at the Center for Open Middleware, which was a collaboration between the UPM and Banco Santander to improve the bank’s technological platform (services and infrastructure). His work was focused on applying machine learning mechanism to process streaming data to predict infrastructure and services failures. Since 2015, he has been an assistant professor at UPM’s Escuela Técnica Superior de Ingeniería y Sistemas de Telecomunicación in the Telematics and Electronics Department where he continues carrying out his research area in machine learning and big data.”
Javier Andion (j.andion@upm.es) is pursuing his Ph.D. in telecommunications engineering at Universidad Politecnica de Madrid. His area of investigation is the proactive detection of failures in computer networks. He joined his current research group in 2015 where his work has been centered in data analytics, software architecture and machine learning in distributed systems
Venue: Eng 209 (Engineering Building)
Time and Date: 1 to 2 PM, Thursday, 25th July 2019.
Please keep the dates on your calendar.
Any comments, advice would be appreciated.