beta-VAE: Learning Basic Visual Concepts with a Constrained Variational Framework
Citations
2,827 citations
2,704 citations
Cites background from "beta-VAE: Learning Basic Visual Con..."
...(12), our simplified objective (14) is a weighted variational bound that emphasizes different aspects of reconstructions that θ must perform [16, 20]....
[...]
1,874 citations
1,612 citations
1,602 citations
Additional excerpts
...Representation Disentanglement [113, 279, 335, 336, 337, 338, 339, 340, 341, 342]...
[...]
References
111,197 citations
"beta-VAE: Learning Basic Visual Con..." refers methods in this paper
...To help quantify the differences, we develop a new measure of disentanglement and show that β-VAE also significantly outperforms all our baselines on this measure (ICA, PCA, VAE by Kingma & Ba (2014), DC-IGN by Kulkarni et al. (2015), and InfoGAN by Chen et al. (2016))....
[...]
38,211 citations
28,898 citations
11,201 citations
"beta-VAE: Learning Basic Visual Con..." refers background in this paper
...It has been suggested that learning a disentangled representation of the data generative factors can be useful for a large variety of tasks and domains (Bengio et al., 2013)....
[...]
...Having a representation that is well suited to the particular task and data domain can significantly improve the learning success and robustness of the chosen model (Bengio et al., 2013)....
[...]
...Learning a disentangled posterior distribution of the generative factors of the observed unsupervised sensory input is a major challenge in AI research (Bengio et al., 2013; Lake et al., 2016)....
[...]
...A disentangled representation can be defined as one where single latent units are sensitive to changes in single generative factors, while being relatively invariant to changes in other factors (Bengio et al., 2013)....
[...]
7,244 citations