New work out on arXiv! Reverse engineering recurrent networks for sentiment classification reveals line attractor dynamics (https://arxiv.org/abs/1906.10720 ), with fantastic co-authors @ItsNeuronal, @MattGolub_Neuro, @SuryaGanguli and @SussilloDavid. #tweetprint summary below!
(1/4)
-
-
Also worth pointing out that these line attractor dynamics were common across LSTMs, GRUs, and vanilla RNNs, which is perhaps surprising given the differing functional forms of these RNN architectures.
Hvala. Twitter će to iskoristiti za poboljšanje vaše vremenske crte. PoništiPoništi
-
Čini se da učitavanje traje već neko vrijeme.
Twitter je možda preopterećen ili ima kratkotrajnih poteškoća u radu. Pokušajte ponovno ili potražite dodatne informacije u odjeljku Status Twittera.


