Invited talk by Dr. Jiaul Paik on High Quality Search- Ranking Models and System Design

Invited talk by Dr. Jiaul Paik on High Quality Search: Ranking Models and System Design

Title of the talk: High Quality Search: Ranking Models and System Design
Speaker: Dr. Jiaul Paik
Time: August 26th (Wednesday) from 3:00 pm - 4:00 pm
Venue: Room 121
Host Faculty: Dr. Manish SIngh

ABSTRACT:

In this talk, I will describe my recent work on ranking models as well as the design of system that will facilitate large scale data processing. I will start by pinpointing the significant limitations of the ranking functions widely used for decades and then describe the development of my work that addresses the problem effectively. To that end, I will present a model that is guided by information theoretic principle.

Although, the existing models are known to be reasonably effective in controlled setup with smaller data set, their brittleness are strikingly evident on very large modern datasets with millions of documents. The next part of my talk will introduce yet another probabilistic ranking model based on extreme value theory that gives state of the art results by outperforming the existing probabilistic models by a significantly large margin. Finally, the last part of my talk will focus on system design for retrieving high quality information from massive amount of data. Specifically, I will talk about my ongoing work on massively parallel text processing algorithms that are built on SPARK, a distributed framework for large scale data processing. I will show a case study where a parallel algorithm executed on a scale-up architecture with significantly lesser computing resources outperforms a powerful scale-out architecture with hundreds of cores and terabytes of memory.

SPEAKER BIO:

I am a postdoctoral researcher at the University of Maryland, College Park, USA (http://www.umiacs.umd.edu/~jiaul/). I work on web search, information retrieval and “big data”, with a particular focus on system building and large-scale distributed algorithms for text processing. Before moving to University of Maryland, I was a visiting scientist at Indian Statistical Institute, Kolkata from where I received my PhD in Computer Science in 2013.

Dates:
Wednesday, August 26, 2015 - 15:00