Upgrade to Pro — share decks privately, control downloads, hide ads and more …

MixPoet

Sponsored · Your Podcast. Everywhere. Effortlessly. Share. Educate. Inspire. Entertain. You do you. We'll handle the rest.

 MixPoet

Avatar for Zhang Yixiao

Zhang Yixiao

April 30, 2020
Tweet

More Decks by Zhang Yixiao

Other Decks in Research

Transcript

  1. 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
  2. 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
  3. 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
  4. 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)
  5. 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。
  6. 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。
  7. Method: Latent Space Mixture • How to learn mixed latent

    space? • For Isotropic Gaussian Space:
  8. 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
  9. Experiments • Baseline: • Ground Truth • C-VAE • USPG

    • MRL: SOTA • fBasic, 监督学习模型
  10. 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