Ehsan Ebrahimzadeh

PhD Candidate

From “Finding a Needle in a Haystack” to “Chasing Vaccination Exemption in Social Media”

  Computer Engineering Department, Floor 1, Class 102
  Wednesday, 30 December 2015
  16:45 - 17:45

Abstract

In the first part of the talk I will be focusing on sequential decision making problems on temporal data. In modern data analytics, there is a growing interest, specially due to constant generation and collection of data over time, in problems where a short term decision or an abrupt response is demanded upon observing new data points. Real-time health monitoring, online advertisement placement, web ranking, anomaly detection, and online fault detection are representative application grounds encompassed within this area, which is usually referred to as online learning and sequential decision making.

After introducing a general framework for online learning and sequential analysis problems, we spend more time on a problem in the context of sequential hypothesis testing, known as change point detection. Specifically, we will look into the problem of detecting an abrupt transient change in the statistical behavior of a random sequence.
If time permits, I will deviate from the somehow theoretical treat of the online learning to a more data-oriented theme of problems, to be unified under “culture analytics”. I will be focusing on a narrative discovery problem from an unstructured corpus of text, an attempt to make connections between data sciences and humanities. Along the way, we will explore some basic tools in hands for analyzing textual data.

Bio

I’m a PhD candidate in the Electrical Engineering Department of UCLA, where I’m simultaneously working towards my Ms degree in Applied Mathematics. Broadly speaking, I’m interested in Machine Learning, Statistics, Applied Probability, and Data Analytics. Before joining UCLA in 2013, I received my MASc degree in Electrical Engineering from University of Waterloo, and my BSc degrees in Mathematics and Electrical Engineering from Isfahan University of Technology.