makrai / toread

Papers I would like to read
0 stars 0 forks source link

Socher Coattention kvázi-rekurrens Many-Task #1

Open makrai opened 6 years ago

makrai commented 6 years ago

Socher ... beszámolt a csapata néhány friss eredményéről, amit reménytelinek látok ebből a szempontból. Ezeket a modelleket még nem tanulmányoztam alaposan, ezért inkább Socher oldalát ajánlom az olvasó figyelmébe: Dynamic Coattention Networks 4 QA, az RNN és a CNN bizonyos tulajdonságait egyesítő kvázi-rekurrens ideghálók, A Deep Reinforced Model for Abstractive Summarization és A Joint Many-Task Model

makrai commented 6 years ago

A Joint Many-Task Model: Growing a Neural Network for Multiple NLP Tasks, 
Kazuma Hashimoto, Caiming Xiong, Yoshimasa Tsuruoka, Richard Socher
EMNLP 2017 + blog post

A Deep Reinforced Model for Abstractive Summarization, 
Romain Paulus, Caiming Xiong, Richard Socher
[ pdf, blog post, Press: Forbes, MIT Tech Review, TechCrunch ]

Dynamic Coattention Networks For Question Answering, 
Caiming Xiong, Victor Zhong, Richard Socher
ICLR 2017+ blog post, Best model on Stanford Question Answering Dataset (At submission) ]

Quasi-Recurrent Neural Networks, 
James Bradbury, Stephen Merity, Caiming Xiong, Richard Socher
ICLR 2017 + blog post