Link
Embed
Share
Beginning
This slide
Copy link URL
Copy link URL
Copy iframe embed code
Copy iframe embed code
Copy javascript embed code
Copy javascript embed code
Share
Tweet
Share
Tweet
Slide 1
Slide 1 text
Decomposable Neural Paraphrase Generation
Slide 2
Slide 2 text
https://arxiv.org/abs/1906.09741
Slide 3
Slide 3 text
• • • •
Slide 4
Slide 4 text
• • •
Slide 5
Slide 5 text
•
Slide 6
Slide 6 text
• • •
Slide 7
Slide 7 text
• • •
Slide 8
Slide 8 text
• •
Slide 9
Slide 9 text
• • = [1 , … , ] • = [1 , … , ]
Slide 10
Slide 10 text
• • • ℎ = BiLSTM( ; ℎ−1 , ℎ+1 ) • = LSTM ℎ , −1 ; −1 • = GumbelSoftmax( , )
Slide 11
Slide 11 text
• • = − encoderz (, ) • 1:−1 , = − encoderz , 1:−1
Slide 12
Slide 12 text
• • 1:−1 , = σ 1:−1 , ( |1:−1 , )
Slide 13
Slide 13 text
• 0 , 1 • = LSTM 0 ; 1 ; −1 ; −1 • 1:−1 , = GumbelSoftmax ,
Slide 14
Slide 14 text
• • ∗ = 0 ∗ = 1
Slide 15
Slide 15 text
• • ℒ = σ=1 log 1:−1 , + σ=1 log ∗ + σ=1 log ( ∗ 1:−1 ,
Slide 16
Slide 16 text
• • •
Slide 17
Slide 17 text
• •
Slide 18
Slide 18 text
•
Slide 19
Slide 19 text
• •
Slide 20
Slide 20 text
• •
Slide 21
Slide 21 text
• •
Slide 22
Slide 22 text
• •
Slide 23
Slide 23 text
• •
Slide 24
Slide 24 text
• • From 1(best) to 4(worst)
Slide 25
Slide 25 text
• • • • •