Building the Metadata Module for the Content Creation Pipeline

Overview

The project aims to enhance the management of training materials by leveraging large language models (LLMs) and advanced machine learning techniques to automate content tagging, categorization, and relationship extraction across formats such as videos, PDFs, Word documents, and audio files. At its core, this initiative addresses the challenges of manual metadata tagging, which is often time-consuming and prone to errors. By integrating SpaCy for entity categorization and Meta LLaMA for extracting context-aware relationships, the project significantly improves the accuracy and efficiency of metadata management. The extracted relationships are structured and stored in a graph database to ensure organized and efficient retrieval of training content. This approach improves data integrity, reduces operational complexities, and allows educational institutions to streamline content management processes while providing a scalable framework that can adapt to future needs and technological advancements.

 

Community Benefit

The automated metadata tagging system improves the scalability and efficiency of training content management, allowing Navy personnel to access relevant materials quickly. Reducing manual effort and minimizing errors ensures the timely delivery of training, which is critical for mission readiness. This system enhances operational effectiveness and supports the long-term adaptability of training processes.

 

Team Members

• Mohammad Chowdhury
• Dylan Hawryluk
• Isaam Siddique 
• Akeno Williams
• Sekai Wynn

 

Sponsored By

Dr. Georgios Sklivanitis