PROJECT TITLE: Technical Training Knowledge Architecture
Project completed May 2018
Collaborators:
- PI: Dr. Chien-Chung Chan, Professor of Computer Science, University of Akron
- Co-PI: Dr. Chen Ling, Associate Professor of Mechanical Engineering, University of Akron
- Co-PI: Dr. Shengyong Wang, Assistant Professor of Mechanical Engineering, University of Akron
- Students: Sai Prajeeth Annamgari
- FAA Sponsor: Luis Ramirez, Technical Training (AJI-2)
- FAA Technical Monitor: John Stillings
ABSTRACT:
The FAA’s technical training involves lots of knowledge from a massive amount of data/documents. This project will explore how knowledge-based question answering systems can support effective training, through studying features of the MIT START system, performing a case study using course materials of FAA, and building a prototype system to demonstrate the potential of how question answering systems can support and enhance training.
Expected Project Outcomes:
- Study features of knowledge-based question answering system to support and enhance effective training. (The system to be analyzed could include the START system developed by MIT.)
- Survey of Advanced Distributed Learning (ADL) initiative by Department of Defense to identify best practices of knowledge-based question answering system in supporting and enhancing effective training.
- Demonstrate how knowledge-based question answering system can support training by performing a case study using course materials of FAA air traffic controller training course: En route stage 1 course: 50148.Survey of Advanced Distributed Learning (ADL) initiative by Department of Defense to identify best practices of knowledge-based question answering system in supporting and enhancing effective training.
VALUE OF RESEARCH:
The system will allow users to ask for a subject of interest, and it will identify a list of documents ordered by degree of relevancy to the subject. In addition, it will also identify topics related to the subject ordered by frequency of occurrence. The project will transform existing training course contents into a repository searchable by subjects of interest.
Our project in this phase provides a foundation and a prototype of how to transform training course materials into a searchable repository by modern NLP tools. With the enhancement of AI tools, it will support the development of effective training systems in the digital era.