Our love for cooking, AI, and trying new things inspired us to create a self-learning AI assistant to learn our personal taste in food and suggest a random recipe that we'll love (hopefully). We wanted to share that experience with you all!
What it does
Our AI assistant learns a user's taste in food through ingredients of dishes. After collecting a reasonable dataset of dishes the user likes/dislikes, the AI can select a random recipe from our database that fits in line with the interests the user has on food. The user can keep track of info as time goes on, andThe AI can continue to learn the user's preference becoming more accurate as the user continues to use it for cooking, or even just for fun.
How we built it
Challenges we ran into
One challenge we faced was figuring out how to design the website and serve data/recipes to the users. Additionally, web-scraping a lot of data from BBC foods as well as parsing that data into a format that we could use was a struggle. Most of all, working for the first time on AI was an experience that gave us a run for our money, and having to traverse through many videos, we barely scratched the surface of what scikit-learn had to offer.
Accomplishments that we're proud of
- Creating a full-functioning website with login and personalized user data.
- Storing all the web scraped data to a postgres database.
- Creating an AI assistant that can serve recipes that are marginally relevant to a user's food interests.
What we learned
We learned to be brave! We should not limit our ideas to our capabilities but learn and try new things.
What's next for TasteAI
- We want to make TasteAI into a public website for all people to use.
- We hope to gain a community who will contribute their own recipes, expanding upon the dataset we have to train our AI assistant.
- Most of all, we hope that TasteAI will provide new recipes that people can try, and know taste good.