NLP Highlights cover image

107 - Multi-Modal Transformers, with Hao Tan and Mohit Bansal

NLP Highlights

00:00

Cosmodaligrams

Thea, thanks the lord for this interesting dis on i had found chatting with you. The thing i want to look at is that whether we could bild the cosmodaligram presentation with unparallel data. So sort of like the non parallel tala that exists in much biggar quantities. And then goes back to some of mat's points about like the latin alignment learning versus some noisy initial allignmentso all of that. That's really interesting things to think about along those lines.

Transcript
Play full episode

The AI-powered Podcast Player

Save insights by tapping your headphones, chat with episodes, discover the best highlights - and more!
App store bannerPlay store banner
Get the app