Re:Deep learning job postings have collapsed in the past six...
by The_rationalist on Monday August 31, 2020 @07:15PM
(#24336840)
I asked the (at the time) SOTA author to replace its BERT implementation with XLnet. He accepted and won 0.5% of accuracy on constituency parsing
https://github.com/sebastianruder/NLP-progress/blob/master/e... which is actually a huge, it means 5% less errors.
Which tasks remain to be tried? Most actually but an obvious one would be coreference resolution https://github.com/sebastianruder/NLP-progress/blob/master/e...