GRS Competition

1st SoIT Storifying your Graduate Research Competition 2023

(Video)

Neuro-symbolic Logic Reasoning Model for Decision Making Support in Legal Scenarios 

by Xiaoxi Kang

Legal language can be confusing and challenging to understand, much like trying to decipher Minion language. Words like "consideration" may have a specific legal meaning that is vastly different from everyday English, leading to confusion for those who are not well-versed in legal jargon. However, with the rapid expansion of Natural Language Processing (NLP) in legal documents, there is hope. Our proposed solution involves training a neuro-symbolic model that can identify legal concepts and provide insightful analysis, supported by relevant information like court cases or statutes. With this trained model, we aim to make legal analysis more accessible and easier to understand by answering users' questions with human-interpretable reasoning traces. So, say goodbye to feeling lost in legal jargon - our solution has got you covered!



(Video)

RevealYourClass: Deciphering Students' Emotions During Online Learning 

by Yohani Ranasinghe

Successful Teaching is not only quality learning content. Emotional interactions with the students are equally important. Rapid transition from physical to online classes, missed “Emotional” piece from the scene. Online learning platforms are getting better at quality content delivery with the advancement of technology. But still we have forgotten the emotions. Can we incorporate Emotion Awareness to online learning? Can machines understand students’ emotions by looking at Facial Expressions, Eye Movement Behavior, and Body Gestures? Is it possible to reveal the emotional state of the online class to the teachers and customize teaching method for a better learning environment in future?

Malaysian English needs Attention

by Mohan Raj Chanthran

I always start my day by reading Malaysian English News. Being a multicultural country, Malaysian English combines elements of standard English with Malay, Chinese, and Indian languages. It is shocking to know Malaysian NLP Landscape is not focusing on Malaysian English. Why not my research be a starting point to build a base for Malaysian English language? So I start by developing an Annotated Malaysian English News Articles dataset, then leverage on the strength of deep learning to extract information from articles. My research can eventually provide an overview, detect redundant news articles, or to construct summary for the news articles.