Upgrade to Pro
— share decks privately, control downloads, hide ads and more …
Speaker Deck
Speaker Deck
PRO
Sign in
Sign up for free
文献紹介: Decomposable Neural Paraphrase Generation
Yumeto Inaoka
July 23, 2019
Research
0
77
文献紹介: Decomposable Neural Paraphrase Generation
2019/07/23の文献紹介で発表
Yumeto Inaoka
July 23, 2019
Tweet
Share
More Decks by Yumeto Inaoka
See All by Yumeto Inaoka
文献紹介: Quantity doesn’t buy quality syntax with neural language models
yumeto
1
39
文献紹介: Open Domain Web Keyphrase Extraction Beyond Language Modeling
yumeto
0
65
文献紹介: Self-Supervised_Neural_Machine_Translation
yumeto
0
51
文献紹介: Comparing and Developing Tools to Measure the Readability of Domain-Specific Texts
yumeto
0
48
文献紹介: PAWS: Paraphrase Adversaries from Word Scrambling
yumeto
0
27
文献紹介: Beyond BLEU: Training Neural Machine Translation with Semantic Similarity
yumeto
0
95
文献紹介: EditNTS: An Neural Programmer-Interpreter Model for Sentence Simplification through Explicit Editing
yumeto
0
130
文献紹介: Analyzing the Limitations of Cross-lingual Word Embedding Mappings
yumeto
0
86
文献紹介: Similarity-Based Reconstruction Loss for Meaning Representation
yumeto
1
70
Other Decks in Research
See All in Research
時間情報表現抽出とルールベース解析器のこれから / Temporal Expression Analysis in Japanese and Future of Rule-based Approach
yag_ays
PRO
0
750
#osc22on 中国オープンソースムーブメントの盛り上がりと 中国最大のオープンソースアライアンス「開源社」の活動紹介
takasumasakazu
0
150
素人発想玄人実行2.0
hf149
0
1.2k
ヒトをめぐるデータサイエンスにおけるAIの最新動向.pdf
tdailab
0
170
データサイエンティストと博士の専門性
mtakano
1
130
計算情報学研究室 (数理情報学第7研究室)紹介スライド
ssakaue
0
120
第8回チャンピオンズミーティング・サジタリウス杯決勝集計 / Umamusume Sagittarius 2021 Final
kitachan_black
0
1.1k
データに基づくレヴィ流行語大賞2021
levii
0
230
SketchODE: Learning neural sketch representation in continuous time
dasayan05
0
110
AIOps研究録―SREのための システム障害の自動原因診断 / SRE NEXT 2022
yuukit
5
1.2k
第1回『帰ってきたウルトラマン』勉強会〜ウルトラマンが帰ってきた〜
hirokundayon
0
540
GDPナウキャスティング・webアプリ「NowcastingR」の概要
secondapunta
0
230
Featured
See All Featured
Design by the Numbers
sachag
271
17k
Six Lessons from altMBA
skipperchong
14
1.3k
Atom: Resistance is Futile
akmur
255
20k
Java REST API Framework Comparison - PWX 2021
mraible
PRO
11
4.6k
Put a Button on it: Removing Barriers to Going Fast.
kastner
56
2.3k
Designing with Data
zakiwarfel
91
3.9k
The Cult of Friendly URLs
andyhume
68
4.7k
What the flash - Photography Introduction
edds
61
9.9k
How GitHub Uses GitHub to Build GitHub
holman
465
280k
BBQ
matthewcrist
74
7.9k
Imperfection Machines: The Place of Print at Facebook
scottboms
253
11k
Design and Strategy: How to Deal with People Who Don’t "Get" Design
morganepeng
103
16k
Transcript
Decomposable Neural Paraphrase Generation
https://arxiv.org/abs/1906.09741
• • • •
• • •
•
• • •
• • •
• •
• • = [1 , … , ] • =
[1 , … , ]
• • • ℎ = BiLSTM( ; ℎ−1 , ℎ+1
) • = LSTM ℎ , −1 ; −1 • = GumbelSoftmax( , )
• • = − encoderz (, ) • 1:−1 ,
= − encoderz , 1:−1
• • 1:−1 , = σ 1:−1 , ( |1:−1
, )
• 0 , 1 • = LSTM 0 ; 1
; −1 ; −1 • 1:−1 , = GumbelSoftmax ,
• • ∗ = 0 ∗ = 1
• • ℒ = σ=1 log 1:−1 , + σ=1
log ∗ + σ=1 log ( ∗ 1:−1 ,
• • •
• •
•
• •
• •
• •
• •
• •
• • From 1(best) to 4(worst)
• • • • •