The aim of this project is to create necessary infrastructure so that machine learning algorithms can be efficiently trained and prediction can be done effectively. The project also requires implementing code classifier which will allow creators to train a classifier for code interaction input. This will allow Oppia to give feedback more efficiently whenever an input fails to classify in any of the hard rules specified by creator, rather than giving default outcome feedback.
Basically there are 3 things to do in this project:
- Create necessary infrastructure for training of machine learning algorithms.
- Implement the code classifier algorithm.
- Enable the machine learning facility in Oppia so that creator can use it.
The infrastructure to be used by Oppia for training of algorithm is called Oppia-ml. It provides Machine learning - as - a - service infrastructure for Oppia. Moreover prediction should be done in the frontend, in users computer so that there is no need of internet connection for giving feedback to the user. The outcome of this project is complete integration of machine learning framework and make it available to end user in production environment.