
Augmenting Named Entity Recognition and Dependency Parsing in the SpaCy NLP Pipeline for Enhanced Mathematical Instruction in Educational Speech-to-Text Systems
This project primarily focuses on augmenting the spaCy NLP pipeline for improved interpretation of complex spoken mathematical expressions. To demonstrate this, the researchers developed TalkBoard, a virtual whiteboard. TalkBoard visualizes speech-to-text mathematical input and performs operations, thereby addressing real-world challenges such as difficulty writing equations in online class or tedious derivations and integrations in educational settings. Currently, TalkBoard covers derivatives and indefinite integrals. It integrates the researchers' fine-tuned DistilBERT model with their enhanced spaCy pipeline for command understanding while utilizing SymPy for computations.
Group Members
Mentor
Topics
AVP
Gallery






