• Corpus ID: 224818149

Probabilistic Numeric Convolutional Neural Networks

  title={Probabilistic Numeric Convolutional Neural Networks},
  author={Marc Finzi and Roberto Bondesan and Max Welling},
Continuous input signals like images and time series that are irregularly sampled or have missing values are challenging for existing deep learning methods. Coherently defined feature representations must depend on the values in unobserved regions of the input. Drawing from the work in probabilistic numerics, we propose Probabilistic Numeric Convolutional Neural Networks which represent features as Gaussian processes (GPs), providing a probabilistic description of discretization error. We then… 

Figures and Tables from this paper

PDE-based Group Equivariant Convolutional Neural Networks

A PDE-based framework that generalizes Group equivariant Convolutional Neural Networks (G-CNNs) and solves the PDE of interest by a combination of linear group convolution and nonlinear morphological group convolutions with analytic kernel approximations that underpin with formal theorems.

Steerable Partial Differential Operators for Equivariant Neural Networks

This work derives a G-steerability constraint that completely characterizes when a PDO between feature vector fields is equivariant, for arbitrary symmetry groups G, and develops a framework forEquivariant maps based on Schwartz distributions that unifies classical convolutions and differential operators and gives insight about the relation between the two.

PolyNet: Polynomial Neural Network for 3D Shape Recognition with PolyShape Representation

A DNN-based method (PolyNet) and a specific polygon mesh representation (PolyShape) with a multi-resolution structure that demonstrates the strength and the advantages of PolyNet on both 3D shape classification and retrieval tasks compared to existingpolygon mesh-based methods and its superiority in classifying graph representations of images.

Graph-Coupled Oscillator Networks

It is proved that GraphCON mitigates the exploding and vanishing gradients problem to facilitate training of deep multi-layer GNNs and offers competitive performance with respect to the state-of-the-art on a variety of graph-based learning tasks.

The Hintons in your Neural Network: a Quantum Field Theory View of Deep Learning

In this work we develop a quantum field theory formalism for deep learning, where input signals are encoded in Gaussian states, a generalization of Gaussian processes which encode the agent’s

Algorithmic Differentiation for Automated Modeling of Machine Learned Force Fields

This paradigmatic approach enables not only a versatile usage of novel representations and the efficient computation of larger systems—all of high value to the FF community—but also the simple inclusion of further physical knowledge, such as higher-order information, even beyond the presented FF domain.

A Practical Method for Constructing Equivariant Multilayer Perceptrons for Arbitrary Matrix Groups

This work provides a completely general algorithm for solving for the equivariant layers of matrix groups and constructs multilayer perceptrons equivariants to multiple groups that have never been tackled before, including the Rubik’s cube group.

Equivariant Deep Learning via Morphological and Linear Scale Space PDEs on the Space of Positions and Orientations

A key isomorphism between linear and morphological scale spaces via the Fourier-Cramér transform maps linear α-stable Lévy processes to Bellman processes and is exploited in PDE-G-CNNs, which generalize Group equivariant Convolutional Neural Networks.



Deep Parametric Continuous Convolutional Neural Networks

The key idea is to exploit parameterized kernel functions that span the full continuous vector space, which allows us to learn over arbitrary data structures as long as their support relationship is computable.

Deep Gaussian Processes with Convolutional Kernels

Convolutional DGP (CDGP) models are developed which effectively capture image level features through the use of convolution kernels, therefore opening up the way for applying DGPs to computer vision tasks.

SplineCNN: Fast Geometric Deep Learning with Continuous B-Spline Kernels

This work presents Spline-based Convolutional Neural Networks (SplineCNNs), a variant of deep neural networks for irregular structured and geometric input, e.g., graphs or meshes, that is a generalization of the traditional CNN convolution operator by using continuous kernel functions parametrized by a fixed number of trainable weights.

Lightweight Probabilistic Deep Networks

  • Jochen GastS. Roth
  • Computer Science
    2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition
  • 2018
This paper proposes probabilistic output layers for classification and regression that require only minimal changes to existing networks and shows that activation uncertainties can be propagated in a practical fashion through the entire network, again with minor changes.

Learning from Irregularly-Sampled Time Series: A Missing Data Perspective

An encoder-decoder framework for learning from generic indexed sequences based on variational autoencoders and generative adversarial networks is introduced and continuous convolutional layers that can efficiently interface with existing neural network architectures are introduced.

Interpolation-Prediction Networks for Irregularly Sampled Time Series

A new deep learning architecture based on the use of a semi-parametric interpolation network followed by the application of a prediction network for addressing the problem of supervised learning with sparse and irregularly sampled multivariate time series is presented.

Multi-Time Attention Networks for Irregularly Sampled Time Series

This work is motivated by the analysis of physiological time series data in electronic health records, which are sparse, irregularly sampled, and multivariate, and proposes a new deep learning framework that is called Multi-Time Attention Networks.

Oriented Response Networks

Over multiple state-of-the-art DCNN architectures, such as VGG, ResNet, and STN, it is consistently observed that replacing regular filters with the proposed ARFs leads to significant reduction in the number of network parameters and improvement in classification performance.

GP-VAE: Deep Probabilistic Time Series Imputation

This work proposes a new deep sequential latent variable model for dimensionality reduction and data imputation of multivariate time series from the domains of computer vision and healthcare, and demonstrates that this approach outperforms several classical and deep learning-based data imputations methods on high-dimensional data.

Set Functions for Time Series

This paper proposes a novel approach for classifying irregularly-sampled time series with unaligned measurements, focusing on high scalability and data efficiency, and is based on recent advances in differentiable set function learning, extremely parallelizable with a beneficial memory footprint.