American University of Beirut (AUB)
Back to all projects
Ongoing
Robust Solutions: Where Black and White Swans are Leveled
About The Project Advisors

Objectives

Would a machine learning algorithm perform better if it were built to treat the highly improbable as fairly probable? Would a stock market crash be prevented or alleviated if the stock crash detection algorithm accounted for extreme events? The answer is yes to both questions as suggested by various recent works. Moreover, when dealing with empirical data, how can a prediction --or an outcome-- be deemed accurate when the used performance-measure tools are intimately tied to the presumption of some “normality” in the observed data, while the underlying laws deviate from the standard?

This project is about developing methodologies to both practical and theoretical problems for which there is sufficient evidence that the "normality" assumption does not hold. More specifically, we propose to design, implement and deploy a complete framework of robust (algorithmic) solutions for a very large set of statistical problems, whenever the empirical data suggest that the underlying uncertainty is "heavy-tailed” in nature. This is also applicable whenever a mathematical model and/or historical data suggest heavy-tailed statistics.

The applications are numerous and span most scientific disciplines such as engineering, finance, economics, computer science, physics, astronomy, medicine, etc.
Methods and Technologies
The desired objectives are reacheable with a complete toolbox for detection and estimation tailored heavy-tailed data. The mathematical background and the theory developement are well in progress and are in one aspect "classical" with parametric data processing, and "contemporary" with statistical tools within the area of artificial intelligence. Additionally, numerical analysis is expected to be challenging and of critical importance within the computational package that naturally requires algorithmic methods and sharp programming skills.
Academic Majors/Disciplines:
Academic Majors/Disciplines:
  • Within MSFEA: CCE - CSE - MECH - BMEN
  • Outside MSFEA: Economics - Finance - Mathematics

Preferred Skills and Experience

  • Technical and mathematical skills
  • Programming skills — Software packages development
  • Data collection and processing
  • Teamwork, communication, and presentation skills

Advisors

Ibrahim Abou-Faycal

ia14@aub.edu.lb

Office: Bechtel 413, Ext: 3454

Jihad Fahs

jf19@aub.edu.lb

Office: Bechtel 428, Ext.: 3631

LinkedIn

Notice :
The attached photos are generated by the AI image generator Fotor using the following keywords: black swan, highly improbable, rare events, prediction, heavy-tailed distributions, outliers

Join today

Apply to the project today, and join other students and faculty members.

Request Information
Newsletter
Join
© 2024 VIP. All rights reserved.
Designed & Developed by Born Interactive