Applying Machine Learning to Detect Internet Anomalies
Machine learning algorithms have been applied to address a variety of engineering and scientific problems. The Internet has historically been prone to failures and attacks that significantly degrade its performance. To detect such anomalies, efforts are made to create datasets from data collected by initiatives such as RIPE, Route Views, and IODA. Creating these datasets involves processing large amounts of data using distributed computing techniques. The application of machine learning allows the detection of anomalies such as outages, worms, ransomware, and DDoS attacks. Machine learning algorithms can not only automate anomaly detection but also enable the detection of patterns not visible to humans due to their complexity and the volume of the number of features and data points. We will describe the entire journey from obtaining raw data through the creation of datasets to the performance results of models capable of detecting anomalies. Furthermore, essential design definitions of the Python code and models developed to carry out such scientific investigations will be addressed.
Date and Time
Location
Hosts
Registration
- Date: 27 Mar 2024
- Time: 04:00 PM to 05:30 PM
- All times are (UTC-07:00) Pacific Time (US & Canada)
-
Add Event to Calendar
- 10704
- Simon Fraser University
- Burnaby, British Columbia
- Canada
- Building: Applied Science Bulding
- Starts 22 March 2024 12:00 AM
- Ends 27 March 2024 01:30 PM
- All times are (UTC-07:00) Pacific Time (US & Canada)
- No Admission Charge
Speakers
Luiz Felipe Silva Oliveira of Simon Fraser University
Applying Machine Learning to Detect Internet Anomalies
Machine learning algorithms have been applied to address a variety of engineering and scientific problems. The Internet has historically been prone to failures and attacks that significantly degrade its performance. To detect such anomalies, efforts are made to create datasets from data collected by initiatives such as RIPE, Route Views, and IODA. Creating these datasets involves processing large amounts of data using distributed computing techniques. The application of machine learning allows the detection of anomalies such as outages, worms, ransomware, and DDoS attacks. Machine learning algorithms can not only automate anomaly detection but also enable the detection of patterns not visible to humans due to their complexity and the volume of the number of features and data points. We will describe the entire journey from obtaining raw data through the creation of datasets to the performance results of models capable of detecting anomalies. Furthermore, essential design definitions of the Python code and models developed to carry out such scientific investigations will be addressed.
Biography:
Dr. Oliveira is a visiting Scholar at Simon Fraser University, Vancouver, Canada, since 2023. He holds Ph.D. in Computing from the System and Computing Engineering Program at COPPE/UFRJ (2019) and Master in Computing by the Postgraduate Program of the Institute of Computing of the Fluminense Federal University (2013). He graduated in Technology in Computer Systems from Universidade Federal Fluminense (2008). He worked for seven years at the Federal Data Processing Service (SERPRO), where he gained experience developing large computational projects for the federal government. Since 2016, he has been a permanent professor at the Federal Institute of Education, Science and Technology of Rio de Janeiro (IFRJ). He has experience in computer science, machine learning, computer networks, and cloud computing. He worked on over 26 projects funded by various institutions and companies, such as the Presidency of the Republic, the Ministry of Planning, the Federal Data Processing Service, and the State of Rio de Janeiro General Public Attorney. His current research interest is in the application of machine learning techniques for classifying and detecting anomalies in time series.
Email:
Address:Simon Fraser Univeristy , , Burnaby, British Columbia, Canada, V5A 1S6