Uncovering data generative factors is the ultimate goal of disentanglement learning. Although many works proposed disentangling generative models able to uncover the underlying generative factors of a dataset, so far no one was able to uncover OOD generative factors (i.e., factors of variations that are not explicitly shown on the dataset). Moreover, the datasets used to validate these models are synthetically generated using a balanced mixture of some predefined generative factors, implicitly assuming that generative factors are uniformly distributed across the datasets. However, real datasets do not present this property. In this work we analyse the effect of using datasets with unbalanced generative factors, providing qualitative and quantitative results for widely used generative models. Moreover, we propose TC-VAE, a generative model optimized using a lower bound of the joint total correlation between the learned latent representations and the input data. We show that the proposed model is able to uncover OOD generative factors on different datasets and outperforms on average the related baselines in terms of downstream disentanglement metrics.

翻译：揭示数据生成因子是解耦学习的最终目标。尽管许多工作提出了能够揭示数据集潜在生成因子的解耦生成模型，但迄今尚无模型能够揭示分布外（OOD）生成因子（即未在数据集中明确展示的变化因子）。此外，用于验证这些模型的数据集通常是通过某些预定义生成因子的平衡混合而合成生成的，这隐含地假设生成因子在数据集中均匀分布。然而，真实数据集并不具备这一特性。本文分析了使用非平衡生成因子数据集的影响，并为广泛使用的生成模型提供了定性与定量结果。此外，我们提出了TC-VAE，一种通过优化学习到的潜在表示与输入数据之间的联合总相关下界来优化的生成模型。实验表明，所提模型能够揭示不同数据集中的分布外生成因子，并在下游解耦指标上平均优于相关基线方法。