UC Berkeley blog on how RNN's may become obsolete

Part 1 (2018)
1

@jeremy, this blog post looks well referenced. Do you think they may be right? It might be a simpler world without RNN’s.

1 Like
2

It’s too early to say, but it’s possible. Like many people, the researchers in that post have ignored QRNNs. Along with our recent work on super-convergence, it’s no longer the case that RNNs need be slow to train.

2 Likes
3

That is what I thought when I saw it posted on LI that maybe they ignored the latest innovations like those in AWD.