R.I.S.E

Your AI assistant for personalized support in daily life.

Comment

Inspiration

Whether it's juggling tasks, accessing information, or seeking assistance, there's a growing demand for personalized support that adapts to individual needs. The idea of RISE was to employ AI technology to help people with these problems by giving them smart, helpful advice that will improve their well-being, convenience, and productivity in a variety of settings and domains.

What it does

Rise helps users manage tasks, access information, and seek assistance through natural language processing and machine learning algorithms. RISE offers features such as scheduling, answering queries, task management, and emotional support, all tailored to the user's preferences and needs. RISE, which utilizes MongoDB, is able to adjust to changing demands and user preferences over time by learning from interactions.

How we built it

Using Electron with as our front end and Django with mongoDB atlas as our back end we made a full stack webapplication.

The front end uses Electron elements to display a fun and responsive circle on screen. The user can talk to the sphere using speech-to-text and the ai responds back with text-to-speech.

Using a key bind, the application takes a screenshot of your screen and uses text detection to isolate text segments on your screen. These text segments along with the users speech are then passed to the back end.

The back end, written in django handles POST requests from the front end. First it sends the text from the front end to OpenAi's GPT 3.5 turbo model. The model is able to decide whether or not it wants to make a query to our vector database, powered by mongo db atlas. The vector database stores values as vectors using the open AI ada-020 embedder. When the AI requests some data, it embeds the request into a vector and then runs a dot product with the whole database, the top vector that is most similar is then sent to the AI and the cycle is repeated. After the AI is completed, its final reply gets stored in the database for future use. So that the AI can get memory and learn from past actions

Challenges we ran into

Difficulties merging the front end and back end

difficulties creating backend and connecting with database difficulties deploying backend

Accomplishments that we're proud of

Our ability to stick together through challenges The vector Database

What we learned

Electron, Django, Mongo Db,

What's next for R.I.S.E

Adding new features to better suit developers: ability to read and write to files creating a much more extensive database

Artists: AI art tools and object detection to help inspire artists given what they have already drawn to inspire them to create more

Built With

Share this project:

Updates