Basic ChatBot working interface
This project aims to create a positive impact on older people and geriatric patients in our society! And according to the World Economic Forum, the old population is going to increase hugely in India and Japan till 2020. "Certainly, this project has a huge impact on society."
A few months earlier, when I was sitting with my grandparents, I felt an absence of some mental support as everyone in my home is busy in there work like office and all. This situation reminded me to make a chatting device which tries to talk with them about how they are feeling and other new talks like how its to be in an amusement park, what's going in France currently etc.
What it does
It just talks Continuously like a human, and it is not designed for narrow-minded tasks (ex, Siri). It's innovative in his talks, and there is no wake word every time to speak.
How I built it
I have built it using a movie dataset, state of the art RNN model for language model, Text to speech function, Multilingual ability through translation, and many more features will be added in future.
Challenges I ran into
Lack of relevant data so will need web scrapping to accumulate data. Have accurate translation A good voice so that we can talk comfortably
Challenges to solve in future
Getting good accuracy of wake word detection and understanding questions to answer. Having all questions and answers in the native language of the user.
Accomplishments that I'm proud of
All the challenges are solved, and the same is accomplishments also theses will be increased after doing future work.
What I learned
I learned various things like- a simple chatbot interface inside python Encoder decoder attention model, Data preprocessing, Making a translation tool, And working on Text to Speech and Speech to text
What's next for BotsEverywhere
I'll make proper hardware for it and could make a new startup by taking this idea. The same thing can also be done through mobile phones on the App, so I'll also make that in future.