NLP Highlights cover image

75 - Reinforcement / Imitation Learning in NLP, with Hal Daumé III

NLP Highlights

00:00

The Cost of a String Edit Distance

If your loss function is string edit distance, then essentially this big is either going to correspond to an insertion of an extra word or it's gonna count as a swap where you swapped the word hippopotamus with a big. But this does depend on exactly how you set up the costs for the different edit operations. You could say swapping horse with pony is less bad than swapping horse with astrophysics.

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