提出非对抗式生成模型，解决传统方法训练不稳与数据覆盖不全问题。
A Non-Adversarial Approach to Idempotent Generative Modelling
- 用重建损失+IMLE替代对抗训练，避免模式崩溃
- 能更好修复损坏数据并生成贴近真实分布的样本
- 隐式学习数据流形距离场，适合生成与修复任务
恒等生成网络（IGNs）是兼具局部数据流形投影功能的深度生成模型，能在数据流形上保持恒等映射，在流形外则实现幂等操作。但其因包含对抗性目标，导致模式崩溃、模式丢失和训练不稳定，仅部分覆盖数据流形，与生成对抗网络存在类似缺陷。本文提出非对抗式恒等生成网络（NAIGNs），采用重建损失与隐式最大似然估计（IMLE）相结合的损失函数，提升了对损坏数据的恢复能力及生成样本与真实数据分布的贴合度。此外，我们证明了NAIGNs能隐式学习到数据流形的距离场，并构建能量基模型。
原文摘要 · Abstract (English)
Idempotent Generative Networks (IGNs) are deep generative models that also function as local data manifold projectors, mapping arbitrary inputs back onto the manifold. They are trained to act as identity operators on the data and as idempotent operators off the data manifold. However, IGNs suffer from mode collapse, mode dropping, and training instability due to their objectives, which contain adversarial components and can cause the model to cover the data manifold only partially -- an issue shared with generative adversarial networks. We introduce Non-Adversarial Idempotent Generative Networks (NAIGNs) to address these issues. Our loss function combines reconstruction with the non-adversarial generative objective of Implicit Maximum Likelihood Estimation (IMLE). This improves on IGN's ability to restore corrupted data and generate new samples that closely match the data distribution. We moreover demonstrate that NAIGNs implicitly learn the distance field to the data manifold, as well as an energy-based model.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。