ASPHD_CSC Archives

PhD Student


Options: Use Forum View

Use Monospaced Font
Show Text Part by Default
Show All Mail Headers

Message: [<< First] [< Prev] [Next >] [Last >>]
Topic: [<< First] [< Prev] [Next >] [Last >>]
Author: [<< First] [< Prev] [Next >] [Last >>]

Print Reply
Raj Sunderraman <[log in to unmask]>
Reply To:
PhD Student <[log in to unmask]>
Wed, 15 Feb 2017 15:04:07 -0500
text/plain (49 lines)
Name of Speaker:  Dr. M. Shahriar Hossain

Place: CS Conference Room, 755 in 25 PP Building

Time: 11:00 am – noon on Thursday, February 16

Title: Storyboarding to Explore and Analyze Evolving Data


Seventy-five percent of the data created each day, as estimated by IBM, is unstructured and 
coming from heterogeneous sources including text, image, and relevant metadata. These 
unstructured data collections are growing everyday resulting in a massive archive for many 
disciplines. This advent of wide-variety of available data-feeds has strengthen the potential to 
analyze lineage of information from interdisciplinary viewpoints, contribute to the growth of 
domain-specific knowledge, and develop a sense of the future-states of specific concepts. 
However, with great possibility comes great challenges – How can we identify the contextual 
snippets from heterogeneous sources that will constitute a specific type of lineage? Can we 
model the influence structure between conceptual phenomena over time? How will we 
meander through multitude of combinations to discover an implicit genealogical graph? 

In this presentation, we will discuss different representations of unstructured text; alignment 
procedures to map heterogeneous snippets such as images, metadata, and entities detected 
within text content; mechanisms to extract different types of context and how those contexts 
evolve over time for topics of interest; and methods to capture eventual and conceptual 
lineage for social and scientific unstructured data. In the social science domain, we will discuss 
how subevents from different times can be connected as a diffusion of context over a timeline. 
Our discussions on the scientific domains will focus on how scientific concepts influence one 
another and evolve over time.

The methodologies I will describe fall under a wide set of analytical algorithms named 
Storyboarding. The presentation will outline how storyboarding fosters interdisciplinary 
research by incorporating objective functions that consider domain knowledge, and by 
emphasizing on experts’ feedback to provide a better understanding of the evolution of today’s 

About the speaker:

Dr. Mahmud Shahriar Hossain has been an Assistant Professor of the Department of Computer 
Science at the University of Texas at El Paso (UTEP) since 2013. He received his PhD from the 
Department of Computer Science at Virginia Tech in 2012. His research interests lie in the 
areas of Big Data Analytics including data mining and machine learning. Along with the 
theoretical aspects of algorithm design for Big Data Analytics, Dr. Hossain is interested in 
building prototypes to solve data mining problems in different disciplines. Dr. Hossain 
collaborates with researchers from many disciplines of the academia, along with scientists 
from the industry and government sectors. His research spans diverse application areas 
including intelligence analysis, sustainability, mechanical engineering, social network analysis, 
software engineering, and biomedical science.