IEEE Trans Neural Netw Learn Syst. 2022 Aug;33(8):3792-3803. doi: 10.1109/TNNLS.2021.3054480. Epub 2022 Aug 3.
In this article, we focus on decomposing latent representations in generative adversarial networks or learned feature representations in deep autoencoders into semantically controllable factors in a semisupervised manner, without modifying the original trained models. Particularly, we propose factors' decomposer-entangler network (FDEN) that learns to decompose a latent representation into mutually independent factors. Given a latent representation, the proposed framework draws a set of interpretable factors, each aligned to independent factors of variations by minimizing their total correlation in an information-theoretic means. As a plug-in method, we have applied our proposed FDEN to the existing networks of adversarially learned inference and pioneer network and performed computer vision tasks of image-to-image translation in semantic ways, e.g., changing styles, while keeping the identity of a subject, and object classification in a few-shot learning scheme. We have also validated the effectiveness of the proposed method with various ablation studies in the qualitative, quantitative, and statistical examination.
在本文中，我们专注于以半监督的方式将生成对抗网络中的潜在表示或深度自动编码器中的学习特征表示分解为语义上可控制的因素，而无需修改原始训练模型。具体来说，我们提出了因子分解纠缠网络（FDEN），它可以学习将潜在表示分解为相互独立的因子。给定一个潜在表示，所提出的框架通过以信息论的方式最小化它们的总相关性来绘制一组可解释的因子，每个因子与独立的变化因子对齐。作为一种插件方法，我们将我们提出的 FDEN 应用于对抗性学习推理和先驱网络的现有网络，并以语义方式执行图像到图像的翻译等计算机视觉任务，例如，在保持主题和对象分类的身份的情况下改变风格在 few-shot 学习方案中。我们还通过定性、定量和统计检验中的各种消融研究验证了该方法的有效性。
