Upgrade to Pro
— share decks privately, control downloads, hide ads and more …
Speaker Deck
Features
Speaker Deck
PRO
Sign in
Sign up for free
Search
Search
MixPoet
Search
Zhang Yixiao
April 30, 2020
Research
4
370
MixPoet
Zhang Yixiao
April 30, 2020
Tweet
Share
More Decks by Zhang Yixiao
See All by Zhang Yixiao
CoCon
ldzhangyx
0
340
vq-cpc
ldzhangyx
0
340
diora
ldzhangyx
0
240
drummernet
ldzhangyx
0
200
ON-LSTM
ldzhangyx
0
150
Other Decks in Research
See All in Research
Large Vision Language Model (LVLM) に関する最新知見まとめ (Part 1)
onely7
22
4.8k
湯村研究室の紹介2024 / yumulab2024
yumulab
0
350
Weekly AI Agents News!
masatoto
26
35k
CUNY DHI_Lightning Talks_2024
digitalfellow
0
130
Whoisの闇
hirachan
3
160
Zipf 白色化:タイプとトークンの区別がもたらす良質な埋め込み空間と損失関数
eumesy
PRO
8
1k
Weekly AI Agents News! 9月号 論文のアーカイブ
masatoto
1
150
2024/10/30 産総研AIセミナー発表資料
keisuke198619
1
380
Tietovuoto Social Design Agency (SDA) -trollitehtaasta
hponka
0
3k
Weekly AI Agents News! 11月号 プロダクト/ニュースのアーカイブ
masatoto
0
200
メールからの名刺情報抽出におけるLLM活用 / Use of LLM in extracting business card information from e-mails
sansan_randd
2
260
機械学習でヒトの行動を変える
hiromu1996
1
380
Featured
See All Featured
Designing for humans not robots
tammielis
250
25k
VelocityConf: Rendering Performance Case Studies
addyosmani
326
24k
How GitHub (no longer) Works
holman
311
140k
Imperfection Machines: The Place of Print at Facebook
scottboms
266
13k
Scaling GitHub
holman
458
140k
Dealing with People You Can't Stand - Big Design 2015
cassininazir
365
25k
I Don’t Have Time: Getting Over the Fear to Launch Your Podcast
jcasabona
29
2k
Templates, Plugins, & Blocks: Oh My! Creating the theme that thinks of everything
marktimemedia
28
2.1k
実際に使うSQLの書き方 徹底解説 / pgcon21j-tutorial
soudai
169
50k
Agile that works and the tools we love
rasmusluckow
328
21k
Building Your Own Lightsaber
phodgson
103
6.1k
The Art of Programming - Codeland 2020
erikaheidi
53
13k
Transcript
MixPoet: Diverse Poetry Generation via Learning Controllable Mixed Latent Space
ArXiv: 2003.06094v1 Presenter: Yixiao Zhang
Overview • Idea: 诗人经历、历史背景等 => 诗歌风格多样化 • Methods: • semi-supervised
VAE • disentangling latent space to sub-spaces • each sub-space corresponds to one factor conditioning • adversarial training
Introduction • 近年的研究,主要考虑语义连贯、主题相关 • 存在diversity的困扰 • diversity: • 主题间多样性:给定两个topic words,生成不同的诗歌
• 主题内多样性:给定一个topic word,生成不同的诗歌 • * 现有的模型倾向于记住常见pattern
Introduction • 生活经历、历史背景、文学流派 => 影响风格
Introduction • MixPoet: semi-supervised VAE • 将latent space分解为sub-spaces,与影响因子一一对应 • 训练阶段:模型预测无label诗歌的factors
• 测试阶段:指定factor的值,生成风格化的诗歌
Related Work • 诗歌生成模型 (RNNs, Memory Models, etc. ) •
多样性的先前研究: • MRL system: 强化学习,鼓励选用高TF-IDF的词汇 • USPG: 无监督最大化style vector和诗歌的mutual information
Related Work • VAE文本生成/诗歌生成 • Yang et. al, 2018b: 学习context-conditioned
latent variable • Hu et al. 2017: 对生成的诗歌进行对抗训练,增强topic相关性 • CVAE 对话多样性: Learning Discourse-level Diversity for Neural Dialog Models using Conditional Variational Autoencoders, ACL 2017 • 本文的对抗:在latent space上做对抗训练
Method • topic keyword: mixture empirical distributions: labeled/ unlabeled
Method: Generator • GRU based model • 是length embedding
Method: Semi-supervised C-VAE • 目的是学习 • 引入z • 由于style与semantics耦合 •
不假设y与z的独立性,而是: • 顺序: w => y => z => x (无y label时)
Method: Semi-supervised C-VAE • then for labeled data: • 估计先验
• 和后验 分别使用一个network计算, recon时最小化KL散度。
Method: Semi-supervised C-VAE • labeled data is too limited •
将y看作另一个latent variable • 估计先验 • 和后验 分别使用一个MLP network计算, recon y时最小化KL散度。
Method: Semi-supervised C-VAE • Total Loss:
Method: Latent Space Mixture • 多个factor时的情形: • 独立性假设:
Method: Latent Space Mixture • How to learn mixed latent
space? • For Isotropic Gaussian Space:
Method: Latent Space Mixture • How to learn mixed latent
space? • For Universal Space: 对于condition: ita是噪声,delta是脉冲函数,c是w, y => 从分布中sample出一个值
Method: Latent Space Mixture • 之后使得discriminator区分这两个z • 估计KL散度: • 其中
就是discriminator
Experiments • factors: • 军旅生涯, 乡村生活, 其他 • 时代繁荣, 时代衰落
• => 6种style
Experiments • Baseline: • Ground Truth • C-VAE • USPG
• MRL: SOTA • fBasic, 监督学习模型
Experiments • 多样性,使用Jaccard Similarity指数评价,越低越好 • 诗歌质量:使用Language Model Score(LMS)评价 • 观察:
• 大多数模型倾向生成重复的短语 • MRL与Basic在intra部分只能生成极其相似的诗歌 • C-VAE情况类似
Experiments • Factor Control Results: • 测试生成的诗歌是否与给定因子类别一致
Experiments • 主观实验
Analysis: Style Mixture
Analysis