Towards a Knowledge Graph Driven Intelligent Tutoring System

Open Access
- Author:
- Mungee, Atharva
- Graduate Program:
- Data Analytics
- Degree:
- Master of Science
- Document Type:
- Master Thesis
- Date of Defense:
- March 22, 2021
- Committee Members:
- Youakim Badr, Thesis Advisor/Co-Advisor
Guanghua Qiu, Committee Member
Raghu Sangwan, Committee Member
Colin Neill, Program Head/Chair - Keywords:
- Knowledge Graph
Knowledge Graph Development
Chatbots
Intelligent Tutoring Systems
NLU
Natural Language
Knowledge Representation - Abstract:
- The rise of Massive Open Online Courses(MOOC) has propelled the rise of E-Learning with all-time availability of learning content in recent times. There has been a rise from 1.6 million to 10.3 million registered students on one of the very popular MOOCs, Coursera. The perpetual availability of lessons and courses has devised a “Learn anywhere anytime approach”. This scenario demands a similar perpetual tutor which can help the students solve topic-related doubts of the users while they are learning about the domain and in turn increasing interaction and engagement of the MOOC platform with the learning content. The traditional Intelligent Tutoring Systems(ITS) lack personalization and fail to effectively adapt to newer and sophisticated neural approaches resulting in reduced effectiveness of knowledge acquisition. Alongside this, these traditional ITS often suffer under-supply of knowledge due to a lack of data available about a certain topic which results in a sparsity of knowledge conveyed to the user. Consequently, the efforts and the cost for developing such systems are considerably high, which results in a relatively high complexity to develop such tutoring systems for different domains at scale. Therefore, it is vital to design an authoring tool that publishes a knowledge-rich domain-specific tutoring assistant which can help the user converse about a particular topic to solve queries/doubts related to a topic the student is learning on the MOOC’s platform. This research presents a methodology to develop a knowledge graph-driven intelligent tutoring system which pedagogically utilizes Natural Language Understanding techniques to improve the quality of interaction of the tutoring system and the student. The work also proposes a methodology to develop a domain-specific open knowledge graph that acts as a knowledge base for the tutoring system. Subsequently, the pipelines developed to create a Knowledge Graph and the NLU pipeline can be reused to produce various domain-specific Intelligent Tutoring Systems resulting in reduced efforts for the development of the same. The NLU model achieves accuracies,f-1, recall, and precision values in the range of 90-95% in classifying the intents and entities in a user message while producing a proof-of-concept Tutoring System for the domain of Artificial Intelligence, Machine Learning. The proof-of-concept contains a total of ~190,000 entities in the Knowledge Graph which fall in the category of the mentioned domain. The NLU pipeline and the Knowledge Graph creation pipeline provide an opportunity to ensure efficient development of a knowledge-rich Intelligent Tutoring System which can converse in Natural Language.