Demystifying Inductive Biases for $\beta$-VAE Based Architectures

• URL: http://arxiv.org/abs/2102.06822v1
• Date: Fri, 12 Feb 2021 23:57:20 GMT
• Title: Demystifying Inductive Biases for $\beta$-VAE Based Architectures
• Authors: Dominik Zietlow, Michal Rolinek, Georg Martius
• Abstract summary: We shed light on the inductive bias responsible for the success of VAE-based architectures. We show that in classical datasets the structure of variance, induced by the generating factors, is conveniently aligned with the latent directions fostered by the VAE objective.
• Score: 19.53632220171481
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: The performance of $\beta$-Variational-Autoencoders ($\beta$-VAEs) and their variants on learning semantically meaningful, disentangled representations is unparalleled. On the other hand, there are theoretical arguments suggesting the impossibility of unsupervised disentanglement. In this work, we shed light on the inductive bias responsible for the success of VAE-based architectures. We show that in classical datasets the structure of variance, induced by the generating factors, is conveniently aligned with the latent directions fostered by the VAE objective. This builds the pivotal bias on which the disentangling abilities of VAEs rely. By small, elaborate perturbations of existing datasets, we hide the convenient correlation structure that is easily exploited by a variety of architectures. To demonstrate this, we construct modified versions of standard datasets in which (i) the generative factors are perfectly preserved; (ii) each image undergoes a mild transformation causing a small change of variance; (iii) the leading \textbf{VAE-based disentanglement architectures fail to produce disentangled representations whilst the performance of a non-variational method remains unchanged}. The construction of our modifications is nontrivial and relies on recent progress on mechanistic understanding of $\beta$-VAEs and their connection to PCA. We strengthen that connection by providing additional insights that are of stand-alone interest.

Related papers

• $α$-TCVAE: On the relationship between Disentanglement and Diversity [21.811889512977924]
  In this work, we introduce $alpha$-TCVAE, a variational autoencoder optimized using a novel total correlation (TC) lower bound. We present quantitative analyses that support the idea that disentangled representations lead to better generative capabilities and diversity. Our results demonstrate that $alpha$-TCVAE consistently learns more disentangled representations than baselines and generates more diverse observations.
  arXiv  Detail & Related papers  (2024-11-01T13:50:06Z)
• Diagnosing and Rectifying Fake OOD Invariance: A Restructured Causal Approach [51.012396632595554]
  Invariant representation learning (IRL) encourages the prediction from invariant causal features to labels de-confounded from the environments. Recent theoretical results verified that some causal features recovered by IRLs merely pretend domain-invariantly in the training environments but fail in unseen domains. We develop an approach based on conditional mutual information with respect to RS-SCM, then rigorously rectify the spurious and fake invariant effects.
  arXiv  Detail & Related papers  (2023-12-15T12:58:05Z)
• Disentanglement via Latent Quantization [60.37109712033694]
  In this work, we construct an inductive bias towards encoding to and decoding from an organized latent space. We demonstrate the broad applicability of this approach by adding it to both basic data-re (vanilla autoencoder) and latent-reconstructing (InfoGAN) generative models.
  arXiv  Detail & Related papers  (2023-05-28T06:30:29Z)
• DA-VEGAN: Differentiably Augmenting VAE-GAN for microstructure reconstruction from extremely small data sets [110.60233593474796]
  DA-VEGAN is a model with two central innovations. A $beta$-variational autoencoder is incorporated into a hybrid GAN architecture. A custom differentiable data augmentation scheme is developed specifically for this architecture.
  arXiv  Detail & Related papers  (2023-02-17T08:49:09Z)
• DAPDAG: Domain Adaptation via Perturbed DAG Reconstruction [78.76115370275733]
  We learn an auto-encoder that undertakes inference on population statistics given features and reconstruct a directed acyclic graph (DAG) as an auxiliary task. The underlying DAG structure is assumed invariant among observed variables whose conditional distributions are allowed to vary across domains led by a latent environmental variable $E$. We train the encoder and decoder jointly in an end-to-end manner and conduct experiments on synthetic and real datasets with mixed variables.
  arXiv  Detail & Related papers  (2022-08-02T11:43:03Z)
• Defending Variational Autoencoders from Adversarial Attacks with MCMC [74.36233246536459]
  Variational autoencoders (VAEs) are deep generative models used in various domains. As previous work has shown, one can easily fool VAEs to produce unexpected latent representations and reconstructions for a visually slightly modified input. Here, we examine several objective functions for adversarial attacks construction, suggest metrics assess the model robustness, and propose a solution.
  arXiv  Detail & Related papers  (2022-03-18T13:25:18Z)
• Regularizing Variational Autoencoder with Diversity and Uncertainty Awareness [61.827054365139645]
  Variational Autoencoder (VAE) approximates the posterior of latent variables based on amortized variational inference. We propose an alternative model, DU-VAE, for learning a more Diverse and less Uncertain latent space.
  arXiv  Detail & Related papers  (2021-10-24T07:58:13Z)
• Local Disentanglement in Variational Auto-Encoders Using Jacobian $L_1$ Regularization [21.80539548847009]
  Variational Auto-Encoders (VAEs) and their extensions have been shown to align latent variables with PCA directions. We propose applying an $L_$1 loss to the VAE's generative Jacobian during training to encourage local latent variable alignment.
  arXiv  Detail & Related papers  (2021-06-05T15:40:55Z)
• NestedVAE: Isolating Common Factors via Weak Supervision [45.366986365879505]
  We identify the connection between the task of bias reduction and that of isolating factors common between domains. To isolate the common factors we combine the theory of deep latent variable models with information bottleneck theory. Two outer VAEs with shared weights attempt to reconstruct the input and infer a latent space, whilst a nested VAE attempts to reconstruct the latent representation of one image, from the latent representation of its paired image.
  arXiv  Detail & Related papers  (2020-02-26T15:49:57Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.