Research Idea Ring Talk (Online Via Zoom)

First half hour:

Speaker: Ayush Kumar Shah

Abstract: We want to build a graph-based mathematical formula parser that matches the current state-of-the-art models in accuracy, while having a faster recognition speed. We study the existing state-of-the-art works in visual parsing from computer vision, emphasizing on parsing of notations like mathematical formulas, chemical structures, tables, etc. We critically analyze the strengths and weaknesses of the existing parsing techniques and identify the problems to be tackled to build a better mathematical parser. The mathematical parser, after integration with a formula detection, can be used for indexing math in technical document collections to support math-aware search engines capable of processing queries containing keywords and formulas.

Bio: I am a 1st year Ph.D. student at the Golisano College of Computing and Information Sciences. I have been working in the Document and Pattern Recognition Lab (DPRL), advised by Professor. Richard Zanibbi. I am a part of the MathSeer project (https://www.cs.rit.edu/~dprl/mathseer/), which is a system to make finding mathematical information easier through innovative search engines, interfaces, and algorithms for extracting and recognizing math. Currently, my research is focused on studying state-of-the-art work in visual parsing in computer vision with an emphasis on parsing of notations, including mathematical expressions, chemical structures, etc in order to build a reliable and fast mathematical formula recognition system.

Advisor: Professor. Richard Zanibbi

Second half hour: TBD