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
250
drummernet
ldzhangyx
0
200
ON-LSTM
ldzhangyx
0
160
Other Decks in Research
See All in Research
【NLPコロキウム】Stepwise Alignment for Constrained Language Model Policy Optimization (NeurIPS 2024)
akifumi_wachi
3
540
Leveraging LLMs for Unsupervised Dense Retriever Ranking (SIGIR 2024)
kampersanda
2
310
Segment Any Change
satai
3
220
大規模日本語VLM Asagi-VLMにおける合成データセットの構築とモデル実装
kuehara
5
1k
博士学位論文予備審査 / Scaling Telemetry Workloads in Cloud Applications: Techniques for Instrumentation, Storage, and Mining
yuukit
1
1.7k
The Economics of Platforms 輪読会 第1章
tomonatu8
0
150
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
430
20241115都市交通決起集会 趣旨説明・熊本事例紹介
trafficbrain
0
1.1k
コーパスを丸呑みしたモデルから言語の何がわかるか
eumesy
PRO
8
2.6k
VisFocus: Prompt-Guided Vision Encoders for OCR-Free Dense Document Understanding
sansan_randd
1
470
o1 pro mode の調査レポート
smorce
0
130
Weekly AI Agents News!
masatoto
31
55k
Featured
See All Featured
個人開発の失敗を避けるイケてる考え方 / tips for indie hackers
panda_program
100
18k
How GitHub (no longer) Works
holman
314
140k
Visualization
eitanlees
146
15k
Statistics for Hackers
jakevdp
797
220k
A Modern Web Designer's Workflow
chriscoyier
693
190k
VelocityConf: Rendering Performance Case Studies
addyosmani
328
24k
実際に使うSQLの書き方 徹底解説 / pgcon21j-tutorial
soudai
175
52k
Principles of Awesome APIs and How to Build Them.
keavy
126
17k
Fight the Zombie Pattern Library - RWD Summit 2016
marcelosomers
233
17k
How to Create Impact in a Changing Tech Landscape [PerfNow 2023]
tammyeverts
49
2.3k
Art, The Web, and Tiny UX
lynnandtonic
298
20k
StorybookのUI Testing Handbookを読んだ
zakiyama
28
5.5k
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