Scalable and Equivariant Spherical CNNs by Discrete-Continuous (DISCO) Convolutions

• URL: http://arxiv.org/abs/2209.13603v1
• Date: Tue, 27 Sep 2022 18:00:01 GMT
• Title: Scalable and Equivariant Spherical CNNs by Discrete-Continuous (DISCO) Convolutions
• Authors: Jeremy Ocampo, Matthew A. Price, Jason D. McEwen
• Abstract summary: No existing spherical convolutional neural network (CNN) framework is both computationally scalable and rotationally equivariant. We develop a hybrid discrete-continuous (DISCO) group convolution that is simultaneously equivariant and computationally to high-resolution. For 4k spherical images we realize a saving of $109$ in computational cost and $104$ in memory usage when compared to the most efficient alternative equivariant spherical convolution.
• Score: 5.8808473430456525
• License: http://creativecommons.org/licenses/by-nc-nd/4.0/
• Abstract: No existing spherical convolutional neural network (CNN) framework is both computationally scalable and rotationally equivariant. Continuous approaches capture rotational equivariance but are often prohibitively computationally demanding. Discrete approaches offer more favorable computational performance but at the cost of equivariance. We develop a hybrid discrete-continuous (DISCO) group convolution that is simultaneously equivariant and computationally scalable to high-resolution. While our framework can be applied to any compact group, we specialize to the sphere. Our DISCO spherical convolutions not only exhibit $\text{SO}(3)$ rotational equivariance but also a form of asymptotic $\text{SO}(3)/\text{SO}(2)$ rotational equivariance, which is more desirable for many applications (where $\text{SO}(n)$ is the special orthogonal group representing rotations in $n$-dimensions). Through a sparse tensor implementation we achieve linear scaling in number of pixels on the sphere for both computational cost and memory usage. For 4k spherical images we realize a saving of $10^9$ in computational cost and $10^4$ in memory usage when compared to the most efficient alternative equivariant spherical convolution. We apply the DISCO spherical CNN framework to a number of benchmark dense-prediction problems on the sphere, such as semantic segmentation and depth estimation, on all of which we achieve the state-of-the-art performance.

Related papers

• Fast, Expressive SE$(n)$ Equivariant Networks through Weight-Sharing in Position-Orientation Space [15.495593104596399]
  We formalize the notion of weight sharing in convolutional networks as the sharing of message functions over point-pairs. We develop an efficient equivariant group convolutional network for processing 3D point clouds.
  arXiv  Detail & Related papers  (2023-10-04T17:06:32Z)
• O$n$ Learning Deep O($n$)-Equivariant Hyperspheres [18.010317026027028]
  We propose an approach to learning deep features equivariant under the transformations of $n$D reflections and rotations. Namely, we propose O$(n)$-equivariant neurons with spherical decision surfaces that generalize to any dimension $n$. We experimentally verify our theoretical contributions and find that our approach is superior to the competing methods for O$(n)$-equivariant benchmark datasets.
  arXiv  Detail & Related papers  (2023-05-24T23:04:34Z)
• Equivalence Between SE(3) Equivariant Networks via Steerable Kernels and Group Convolution [90.67482899242093]
  A wide range of techniques have been proposed in recent years for designing neural networks for 3D data that are equivariant under rotation and translation of the input. We provide an in-depth analysis of both methods and their equivalence and relate the two constructions to multiview convolutional networks. We also derive new TFN non-linearities from our equivalence principle and test them on practical benchmark datasets.
  arXiv  Detail & Related papers  (2022-11-29T03:42:11Z)
• PDO-e$\text{S}^\text{2}$CNNs: Partial Differential Operator Based Equivariant Spherical CNNs [77.53203546732664]
  We use partial differential operators to design a spherical equivariant CNN, PDO-e$textStext2$CNN, which is exactly rotation equivariant in the continuous domain. In experiments, PDO-e$textStext2$CNNs show greater parameter efficiency and outperform other spherical CNNs significantly on several tasks.
  arXiv  Detail & Related papers  (2021-04-08T07:54:50Z)
• Rotation-Invariant Autoencoders for Signals on Spheres [10.406659081400354]
  We study the problem of unsupervised learning of rotation-invariant representations for spherical images. In particular, we design an autoencoder architecture consisting of $S2$ and $SO(3)$ convolutional layers. Experiments on multiple datasets demonstrate the usefulness of the learned representations on clustering, retrieval and classification applications.
  arXiv  Detail & Related papers  (2020-12-08T15:15:03Z)
• Efficient Generalized Spherical CNNs [7.819876182082904]
  We present a generalized spherical CNN framework that encompasses various existing approaches and allows them to be leveraged alongside each other. We show that these developments allow the construction of more expressive hybrid models that achieve state-of-the-art accuracy and parameter efficiency on spherical benchmark problems.
  arXiv  Detail & Related papers  (2020-10-09T18:00:05Z)
• PDO-eConvs: Partial Differential Operator Based Equivariant Convolutions [71.60219086238254]
  We deal with the issue from the connection between convolutions and partial differential operators (PDOs) In implementation, we discretize the system using the numerical schemes of PDOs, deriving approximately equivariant convolutions (PDO-eConvs) Experiments on rotated MNIST and natural image classification show that PDO-eConvs perform competitively yet use parameters much more efficiently.
  arXiv  Detail & Related papers  (2020-07-20T18:57:26Z)
• Spin-Weighted Spherical CNNs [58.013031812072356]
  We present a new type of spherical CNN that allows anisotropic filters in an efficient way, without ever leaving the sphere domain. The key idea is to consider spin-weighted spherical functions, which were introduced in physics in the study of gravitational waves. Our method outperforms previous methods on tasks like classification of spherical images, classification of 3D shapes and semantic segmentation of spherical panoramas.
  arXiv  Detail & Related papers  (2020-06-18T17:57:21Z)
• Linear Time Sinkhorn Divergences using Positive Features [51.50788603386766]
  Solving optimal transport with an entropic regularization requires computing a $ntimes n$ kernel matrix that is repeatedly applied to a vector. We propose to use instead ground costs of the form $c(x,y)=-logdotpvarphi(x)varphi(y)$ where $varphi$ is a map from the ground space onto the positive orthant $RRr_+$, with $rll n$.
  arXiv  Detail & Related papers  (2020-06-12T10:21:40Z)
• Robustly Learning any Clusterable Mixture of Gaussians [55.41573600814391]
  We study the efficient learnability of high-dimensional Gaussian mixtures in the adversarial-robust setting. We provide an algorithm that learns the components of an $epsilon$-corrupted $k$-mixture within information theoretically near-optimal error proofs of $tildeO(epsilon)$. Our main technical contribution is a new robust identifiability proof clusters from a Gaussian mixture, which can be captured by the constant-degree Sum of Squares proof system.
  arXiv  Detail & Related papers  (2020-05-13T16:44:12Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.