This idea came into our mind as we usually face this real problem in our student life while preparing and revising notes for semesters exams. The sole concept is based on the Ebbinghaus Forgetting Curve Theory, according to which, the more you revise what you’ve learned, the flatter your forgetting curve gets.
What it does
A Google Action that helps you learn things faster and better than ever before.
How I built it
Used Google Actions as an I/O with user coupled with Google NLP API for some prelim text processing. Actions SDK was linked up to DialogFlow for NLU logic. Server was based off of MEAN stack for scheduling and notifications and is hosted off of Heroku.
Challenges I ran into
The main technical challenge among other challenges was to re-modeled the Ebbinghaus's forgetting curve formula: R = e^(-t/s) and implement an algorithm to compute the data. Challenges: 1)Adjusting the value of S (where S is the “strength” of your memory) and t is the amount of time that has passed. No specific paper was published so we could not get the approximate value of S. 2) Integrating two Rest API server calls to update the backend details of Questions and answers 3) Incorporating Google Action SDK and Dialogflow both.
Accomplishments that I'm proud of
Developing an app which can really help all the students and people with memory based impediments. Correctly re-modeled the equation and create and analyzed datasets to develop real-time visualization report of revision status of the question.
What I learned
Working on postman and REST API Integrating DialogFlow and Google Action Sdk to control Google Assistant App
What's next for iRecallBot
We can incorporate Machine learning algorithms to the proximate value of S as it varies from the brain to brain. Expand the implementation using Google Mini or Alexa Utilizing One-note API to directly feed data as JSON into our database in real time