High-Level Project Summary
There are a large number of training available in the category Earth Observations (EO) and the number of training/resources are increasing as new material is added over time. It is very laborious for any person to scan through the material available on the net and select a specific training/ course that most suits his/her needs.The project presents a Virtual Capacity Building Councillor (gateway) using Natural Language Processing (NLP) to determine the users’ requirements and then matches it with the course learning outcome (CLOs) and objectives of training programs available. Other user requirements like costings, Education level, institution etc are also considered before recommending
Link to Final Project
Link to Project "Demo"
Detailed Project Description
Project Overview
The core of the project is based on NLP engine, which will take input from the user interested in developing his/her capacity in the domain of EO. The input to the system could be text or voice interaction, whichever preferred by the used
The NLP input would consist of answers to the following simple questions
1. What area of the Capacity building are you interested in?
2. What learning objectives do you have in mind?
3. Any specific skill-set that you are interested in?
4. What educational level of training you are looking for?
5. Are you targeting for free training programs?
6. What training time duration are you looking for?
Most of these questions will be optional as the candidate might be undecided in many cases, but providing more information would be better
The preferences of the user will be matched with the training programs/ Capacity building resources available with the NASA resource reservoir. The resource reservoir will be updated regularly from the net with the help of web crawlers, which would be updating new information.
It would be helpful if the new capacity development resources added by the training organizations include a description of specific learning outcomes or value added by the specific resource.
At the end of the course selection process, the participant would also be providing feedback on the issues he/she faced during the selection of the training. This feedback would be used for the learning / services improvement of the virtual councillor.
Technical Details
The front end of the website will use an Avatar to gather information from the user in the form of text or voice. The Avatar would ask pre-determined questions to the user and extract user requirements from the input. NLP will be used for extracting user requirements from the user input.
The different steps during the processing are represented in the processing diagram
Processing Diagram
The back-end processing would include removing of background noise from the voice signal input and voice enhancement treatment. The system can also be trained for detection of slang in the voice input.
Database would be used for storing information on OE courses/ training as well as user requirements for future references.
Space Agency Data
NASA Applied Remote Sensing Training (ARSET) Program
NASA ARSET full catalog of trainings in spreadsheet format with all the relevant data
EO Dashboard
EO College Courses
Hackathon Journey
It was the first time that the team participated in the Hackathon and experienced the learning in a roller-skate manner, at times we thought that we have achieved a strong answer to the problem under study, but the next moment a certain question from the team member would bring us back to step one and we would again start thinking about our fundamentals.
We also learned a lot from informal discussions during the break, from other team members working on different problems as they explained their specific problems and how they were planning to solve their hurdles. The role played by the mentors cannot be overemphasized, as they provided guidance and opened new ways of approaching the problem.
In short, it was a great experience and a good opportunity to make friends
References
Tags
#knowledge management, #environment, #Training, #Earth Observation, #Virtual Assistant