119 - Social NLP, with Diyi Yang
NLP Highlights
00:00
The Importance of Bias Neutralization
We found more errors when it comes to the actual picking the bias the word or like suggesting the appropriate word to replace that. And another kind of shortcut that our models tended to make is it has a tendency to simply remove words instead of finding a good replacement because that's easier. Another aspect that I think we happen to take into account by the issue is to incorporate the aspect of fact checking. That requires us for more sophisticated systems designed for this work.
Transcript
Play full episode
