Tripod：用于解缠表示学习的三种互补归纳偏置
Tripod: Three Complementary Inductive Biases for Disentangled Representation Learning
AI总结：
本文提出Tripod模型，通过整合量化、独立性和最小功能影响三种互补归纳偏置，并加以适应性改进，在四个图像解缠基准上取得最先进性能。
摘要：
归纳偏置在解缠表示学习中至关重要，用于缩小欠定解集的范围。在本工作中，我们考虑为神经网络自编码器赋予文献中三种精选的归纳偏置：通过量化将数据压缩到网格状潜空间、潜变量之间的集体独立性，以及任何潜变量对其他潜变量决定数据生成过程的最小功能影响。原则上，这些归纳偏置是深度互补的：它们分别最直接地指定了潜空间、编码器和解码器的属性。然而在实践中，天真地组合实现这些归纳偏置的现有技术未能产生显著收益。为解决这一问题，我们提出了对这三种技术的适应性改进，以简化学习问题，为关键正则化项配备稳定不变性，并消除退化激励。由此产生的模型Tripod在四个图像解缠基准套件上取得了最先进的结果。我们还验证了Tripod显著优于其天真版本，并且其三条“腿”对于最佳性能都是必要的。
英文摘要：
Inductive biases are crucial in disentangled representation learning for narrowing down an underspecified solution set. In this work, we consider endowing a neural network autoencoder with three select inductive biases from the literature: data compression into a grid-like latent space via quantization, collective independence amongst latents, and minimal functional influence of any latent on how other latents determine data generation. In principle, these inductive biases are deeply complementary: they most directly specify properties of the latent space, encoder, and decoder, respectively. In practice, however, naively combining existing techniques instantiating these inductive biases fails to yield significant benefits. To address this, we propose adaptations to the three techniques that simplify the learning problem, equip key regularization terms with stabilizing invariances, and quash degenerate incentives. The resulting model, Tripod, achieves state-of-the-art results on a suite of four image disentanglement benchmarks. We also verify that Tripod significantly improves upon its naive incarnation and that all three of its "legs" are necessary for best performance.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
