Resources： Disentangled Representation Learning

Introduction

领域解纠缠表示学习（Disentangled Representation Learning）的相关资料：包含论文、tutorial等。

主要的分类： (1) 使用方法： GAN, VAE 等; (2) 监督类型: 无监督、半监督(弱监督，使用 group-paired 的样本)。

Disentangled Representation Learninig

CausalVAE: Structured Causal Disentanglement in Variational Autoencoder. [paper]. [Notes] 引入因果推理。 (CausalVAE)
Controllable Variational Autoencoder. [paper]. [Notes] 引入PI控制器利用反馈动态调节超参数 ${\beta}$ 。 (ControlVAE)
q-VAE for Disentangled Representation Learning and Latent Dynamical Systems. [paper]
An Improved Semi-Supervised VAE for Learning Disentangled Representations.
Disentangled Representation Learning and Generation with Manifold Optimization.
Guided Variational Autoencoder for Disentanglement Learning.
Progressive Learning and Disentanglement of Hierarchical Representations. ( ICLR2020 )

Bayes-Factor-VAE: Hierarchical Bayesian Deep Auto-Encoder Models for Factor Disentanglement. (ICCV2019 Oral) [paper] [slide&video] [codes] [Notes] (BF-VAE)
Relevance Factor VAE: Learning and Identifying Disentangled Factors. [paper] [Notes] (RF-VAE)
Variational Autoencoders and Nonlinear ICA.

Towards a Definition of Disentangled Representations. [paper]
Understanding disentangling in ${\beta}$ -VAE. [paper] [Notes] (AnnealVAE)
Isolating Sources of Disentanglement in VAEs. [paper] [Notes] (ICLR 2018 Oral). ( ${\beta}$ -TCVAE + MIG Metric).
Challenging Common Assumptions in the Unsupervised Learning of Disentangled Representations. [paper] [Notes] (ICML2019 Best Paper)
A framework for the quantitative evaluation of disentangled representations. [paper] (ICLR 2018).
Disentangling by factorising. [paper] [Notes] (ICML 2018).(FactorVAE + FactorVAE Metric)（改进 ${\beta}$ -VAE Metric）
Learning Disentangled Joint Continuous and Discrete Representations (JointVAE). (NIPS2018). [paper] [codes] [Notes]
Hyperprior induced unsupervised disentanglement of latent representations. [paper] [Notes]
Disentangling the independently controllable factors of variation by interacting with the world. [paper]
Structured Disentangled Representations. [paper] [Notes] (HF-VAE)
Disentangling Disentanglement in Variational Autoencoders. [paper] [Notes] (ICML2019)

Variation inference of disentangled latent concepts from unlabeled observations. [paper] [Notes]
Learning independent features with adversarial nets for non-linear ICA. [paper]

InfoGAN: Interpretable Representation Learning by Information Maximizing Generative Adversarial Nets. [paper]
${\beta}$ -VAE: Learning Basic Visual Concepts with a Constrained Variational Framework. [paper] [codes] [Notes]
Disentangling factors of variation in deep representations using adversarial training. [paper]
Adversarial Autoencoders. [paper] [Notes] (AAE)