MSc Thesis Defense of Mrulay Mistry:"GraphWords: Enhancing Word Vectors using Graphs"

 

Date: Monday, April 24, 2023

Time: 3:00 PM to 4:30 PM

Location: Essex Hall, Room 122

 

Reminders: 1. Two-part attendance mandatory (sign-in sheet, QR Code) 2. Arrive 5-10 minutes prior to event starting - LATECOMERS WILL NOT BE ADMITTED. Note that due to demand, if the room has reached capacity, even if you are "early" admission is not guaranteed. 3. Please be respectful of the presenter by NOT knocking on the door for admittance once the door has been closed whether the presentation has begun or not (If the room is at capacity, overflow is not permitted (ie. sitting on floors) as this is a violation of the Fire Safety code). 4. Be respectful of the decision of the advisor/host of the event if you are not given admittance. The School of Computer Science has numerous events occurring soon.

 

 

In the domain of Natural Language Processing (NLP), the representation of words according to their distribution in the vector form is a crucial task. In the representation space, when words that are similar to each other according to human interpretation are placed closer to each other, a notable increase can be observed in the performance and accuracy of NLP tasks. Previous word embedding methods put emphasis on passing the word tokens in an iterative manner through a Neural Language model to capture the context of the words. These methods can capture word relatedness, but only within the given context length. In this thesis, we introduce a method to represent the words in the form of a graph so that their first-order and second-order proximity are preserved, and the relatedness of a word can be captured in a better manner. This graph is then subjected to a vertex (node) embedding method to generate their embedding. After experimenting with the proposed method on multiple text corpora, our findings indicate that this method of word representation outperforms the traditional word-embedding methods by more than 10% in multiple intrinsic and extrinsic tasks.

 

Keywords: Graphs, Natural Language Processing, Node Embedding, Neural Word Embedding

 

Internal Reader: Dr. Hossein Fani   

External Reader: Dr. Reza Nakhaie

Advisor: Dr. Ziad Kobti

Chair: Dr. Kalyani Selvarajah