Network Bending: Expressive Manipulation of Generative Models in Multiple Domains.
Network Bending: Expressive Manipulation of Generative Models in Multiple Domains.
复制标题
DOI：
10.3390/e24010028
复制
发表时间：
2021-12-24
期刊：
影响因子：
--
通讯作者：
Grierson M
中科院分区：
文献类型：
--
作者：
Broad T;Leymarie FF;Grierson M
本文提出了网络弯曲框架，这是一种与深度生成模型进行操作和交互的新方法。我们提出了一套全面的确定性转换，可以作为不同的层插入到训练生成神经网络的计算图中，并在推理过程中应用。此外，我们还提出了一种基于空间激活图分析深度生成模型和聚类特征的新算法。这允许以无监督的方式基于空间相似性将特征分组在一起。这导致对特征集进行有意义的操作，这些特征集对应于生成结果的大量语义上重要特征的生成。我们概述了这个框架，展示了我们在图像和音频领域的深度生成模型上的结果。我们展示了它如何允许对生成过程的语义有意义的方面进行直接操作，以及允许广泛的表达结果。
This paper presents the network bending framework, a new approach for manipulating and interacting with deep generative models. We present a comprehensive set of deterministic transformations that can be inserted as distinct layers into the computational graph of a trained generative neural network and applied during inference. In addition, we present a novel algorithm for analysing the deep generative model and clustering features based on their spatial activation maps. This allows features to be grouped together based on spatial similarity in an unsupervised fashion. This results in the meaningful manipulation of sets of features that correspond to the generation of a broad array of semantically significant features of the generated results. We outline this framework, demonstrating our results on deep generative models for both image and audio domains. We show how it allows for the direct manipulation of semantically meaningful aspects of the generative process as well as allowing for a broad range of expressive outcomes.
影响因子：
6.2
作者：
Bau, David;Strobelt, Hendrik;Torralba, Antonio
通讯作者：
Torralba, Antonio
影响因子：
8.5
作者：
Celebi, M. Emre;Kingravi, Hassan A.;Vela, Patricio A.
通讯作者：
Vela, Patricio A.
影响因子：
0.3
作者：
Broad, Terence;Grierson, Mick
通讯作者：
Grierson, Mick