arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2301.08883cs.LGeess.SP
收藏

学习内隐神经表征的多功能神经过程

Versatile Neural Processes for Learning Implicit Neural Representations

University of Science and Technology of China(中国科学技术大学);Microsoft Research Asia(微软亚洲研究院)

Zongyu Guo, Cuiling Lan, Zhizheng Zhang, Yan Lu, Zhibo Chen

摘要:

将信号表示为由神经网络(a.k.a.近年来,内隐神经表征(INRs)引起了越来越多的关注。神经过程(NPs)对以部分观测(上下文集)为条件的函数上的分布进行建模,为连续函数的快速推理提供了一种实用的解决方案。然而,现有的NP架构遭受劣质复杂信号的建模能力。在本文中,我们提出了一个高效的NP框架称为通用神经过程(VNP),它大大提高了逼近功能的能力。具体来说,我们引入了一个瓶颈编码器,产生更少的信息上下文令牌,减轻高计算成本,同时提供高建模能力。在解码器端,我们分层学习多个全局隐变量,这些隐变量联合建模函数的全局结构和不确定性,使我们的模型能够捕获复杂信号的分布。我们证明了所提出的VNP在涉及1D,2D和3D信号的各种任务上的有效性。特别是,我们的方法在学习准确的INR w.r.t.一个3D场景而无需进一步微调。代码可在https://github.com/ZongyuGuo/Versatile-NP上获得。

英文摘要:

Representing a signal as a continuous function parameterized by neural network (a.k.a. Implicit Neural Representations, INRs) has attracted increasing attention in recent years. Neural Processes (NPs), which model the distributions over functions conditioned on partial observations (context set), provide a practical solution for fast inference of continuous functions. However, existing NP architectures suffer from inferior modeling capability for complex signals. In this paper, we propose an efficient NP framework dubbed Versatile Neural Processes (VNP), which largely increases the capability of approximating functions. Specifically, we introduce a bottleneck encoder that produces fewer and informative context tokens, relieving the high computational cost while providing high modeling capability. At the decoder side, we hierarchically learn multiple global latent variables that jointly model the global structure and the uncertainty of a function, enabling our model to capture the distribution of complex signals. We demonstrate the effectiveness of the proposed VNP on a variety of tasks involving 1D, 2D and 3D signals. Particularly, our method shows promise in learning accurate INRs w.r.t. a 3D scene without further finetuning. Code is available at https://github.com/ZongyuGuo/Versatile-NP .

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。

↑