Generative modeling of convolutional neural networks
From MaRDI portal
Abstract: The convolutional neural networks (CNNs) have proven to be a powerful tool for discriminative learning. Recently researchers have also started to show interest in the generative aspects of CNNs in order to gain a deeper understanding of what they have learned and how to further improve them. This paper investigates generative modeling of CNNs. The main contributions include: (1) We construct a generative model for the CNN in the form of exponential tilting of a reference distribution. (2) We propose a generative gradient for pre-training CNNs by a non-parametric importance sampling scheme, which is fundamentally different from the commonly used discriminative gradient, and yet has the same computational architecture and cost as the latter. (3) We propose a generative visualization method for the CNNs by sampling from an explicit parametric image distribution. The proposed visualization method can directly draw synthetic samples for any given node in a trained CNN by the Hamiltonian Monte Carlo (HMC) algorithm, without resorting to any extra hold-out images. Experiments on the challenging ImageNet benchmark show that the proposed generative gradient pre-training consistently helps improve the performances of CNNs, and the proposed generative visualization method generates meaningful and varied samples of synthetic images from a large-scale deep CNN.
Recommendations
- Maximum likelihood features for generative image models
- GSNs: generative stochastic networks
- Words as a window: using word embeddings to explore the learned representations of convolutional neural networks
- Sparse and deep generalizations of the FRAME model
- A Generative Variational Model for Inverse Problems in Imaging
Cited in
(19)- GANs for generating EFT models
- Generative adversarial networks with decoder-encoder output noises
- A tale of three probabilistic families: discriminative, descriptive, and generative models
- Inducing wavelets into random fields via generative boosting
- scientific article; zbMATH DE number 6395619 (Why is no real title available?)
- scientific article; zbMATH DE number 7625200 (Why is no real title available?)
- Some theoretical properties of GANs
- A predictive-coding network that is both discriminative and generative
- Embedding decision diagrams into generative adversarial networks
- Neural Simpletrons: Learning in the Limit of Few Labels with Directed Generative Networks
- Interpretable Generative Adversarial Networks With Exponential Function
- Co-operative generative adversarial nets
- Maximum likelihood features for generative image models
- Generating Videos Based on Convolutional Recurrent Generative Adversarial Networks
- GSNs: generative stochastic networks
- Words as a window: using word embeddings to explore the learned representations of convolutional neural networks
- Generative model of autoencoders self-learning on images represented by count samples
- Deep visualization based on the spatial pyramid decomposition
- Rectified binary convolutional networks with generative adversarial learning
This page was built for publication: Generative modeling of convolutional neural networks
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1747609)
