Inspiration
Currently AI is booming and it was hard for me to keep up with constant news. So I decided to automate the steps that I'm doing: Go through news platforms and aggregate all relevant AI news.
What it does
It scrapes GitHub, Hackernews and Twitter and filters out AI content. What is unique about it, it deduplicates the news on Twitter.
How we built it
- Using Huggingface Transformers SetFit for classification
- Using Sentence-Embeddings for clustering
- Using Lexrank for scoring
Challenges we ran into
- Getting data from the Twitter API
- Labelling the data
Accomplishments that we're proud of
Check it out: http://news.mioses.com
What we learned
- SetFit for sentence classifcation
- How to use Google App Script to prototype a backend ## What's next for Mioses News - HN for ML Let's make this the go to platform to catch up
Log in or sign up for Devpost to join the conversation.