[month] [year]

Muthyala Anurag Reddy

Muthyala Anurag Reddy supervised by Prof. Vikram Pudi received his Master of Science – Dual Degree  in Computer Science and Engineering (CSE). Here’s a summary of his research work on Hierarchical Curriculum Generation using Topic Modeling and Document Sequence Generation:

In the past, curriculum generation has been an elaborate process requiring a team of experts. The resulting curriculum is suitable for all the students in a discipline and is usually limited to a single discipline. The first part of the thesis explores using natural language processing (NLP) techniques for text generation and their application in wandering curriculum generation from a given list of documents. Specifically, we investigate the use of Wikipedia as an additional external knowledge source for generating information and concise academic curricula. We present a detailed analysis of our custom NLP technique, and its applicability in curriculum generation. We evaluate the effectiveness of our approach by conducting experiments on a large date of academic curriculum content, especially content (NCERT textbook) that is widely used in India in most academic settings. Our results allow that our approach significantly outperforms existing unsupervised approaches in generating informative and relevant curriculum content.
In the second part of this thesis, we address the crucial challenge of organizing the modules and topics generated through natural language processing into a coherent and logical sequence. While the initial phase of our research has successfully demonstrated the capability of NLP techniques, the effectiveness of such content heavily relies on the arrangement and flow of these components. The problem of ordering these modules and topics to create a structured and meaningful academic curriculum is complex and essential task. In this part of our study, we delve into the realm of document arrangement using a new sequence generation model. In today’s information-rich environment, effective structuring and dressing digital documents is of utmost importance. The system examines semantic and graph patterns within documents to create well-organized and contextually relevant sequences. By conducting a thorough assessment of current methods and introducing a new, sequence generation model, this research aims to improve information retrieval, reduce the cognitive burden, and modernize knowledge management.

 September 2024