Into the Unknown

Inspiration

Memories and experiences play a fundamental role in shaping human identity. With the advancing technology, capturing and storing memories has become easier. Utilizing Natural Language Processing (NLP) techniques, we propose a novel approach to construct personalized chatbots. By extracting and processing user memories through digital diaries, we intend to train chatbots to emulate users' experiences and knowledge. Others can then use the platform to “talk” to the user through this bot, hold conversations ask questions, and interact with the bot as if they were talking to the user themselves.

What it does

It actually can store their diary entries, and the quotes(i.e. on what they believe in), and people can view and edit their quotes and diaries, and then they have an option of a chatbot whee they can chat with their past self, or they can chat with someone else bot made of their's memories and quotes they written

How we built it

We have used the next.js app to build a website, and we used Firebase to store the diary entries and blogs and the unique users id's between the shared links we have used Firebase for the authentication purposes and Open AI key for the processing of the text we collect from users

Challenges we ran into

The challenges we ran into were: Was not able to create the own model for time constraint I was not able to find any pre-trained models that work well for free Getting the OpenAI key

Accomplishments that we're proud of

We successfully build the website and deploy it We successfully make the bot ready, and it is answering based on the memories and the quotes of that user

What we learned

We learned on how hackathons work and got different insights and types of projects we can build

What's Next for Into the Unknown

We couldn't build the model to be so personalized with the specific person, but it will be familiar to all the persons.

We will give an option for the user to upload the whatsapp export chat, and then the model will learn the data on how the person usually speaks with the other
We will fine-tune the model for the Q&A and DocumentQ&A based on the sentiment activity for more personalized experience