Introspective Classification with Convolutional Nets

Abstract

We propose introspective convolutional networks (ICN) that emphasize the importance of having convolutional neural networks empowered with generative capabilities. We employ a reclassification-by-synthesis algorithm to perform training using a formulation stemmed from the Bayes theory. Our ICN tries to iteratively: (1) synthesize pseudo-negative samples; and (2) enhance itself by improving the classification. The single CNN classifier learned is at the same time generative --- being able to directly synthesize new samples within its own discriminative model. We conduct experiments on benchmark datasets including MNIST, CIFAR-10, and SVHN using state-of-the-art CNN architectures, and observe improved classification results.

12 pages, 3 figure

Cited by in corpus (14)

• Hybrid Discriminative-Generative Training via Contrastive Learning
• Attended Temperature Scaling: A Practical Approach for Calibrating Deep Neural Networks
• Learning Energy-Based Models by Diffusion Recovery Likelihood
• Reusing Discriminators for Encoding: Towards Unsupervised Image-to-Image Translation
• Learning to Generate Synthetic Data via Compositing
• Learning Latent Space Energy-Based Prior Model
• Flow Contrastive Estimation of Energy-Based Models
• Dual Contradistinctive Generative Autoencoder
• $(1+\epsilon )$-class Classification: an Anomaly Detection Method for Highly Imbalanced or Incomplete Data Sets
• Learning Energy-Based Models as Generative ConvNets via Multi-grid Modeling and Sampling
• Learning Energy-based Spatial-Temporal Generative ConvNets for Dynamic Patterns
• Representation Learning: A Statistical Perspective
• Resisting Large Data Variations via Introspective Transformation Network
• A Novel Unsupervised Post-Processing Calibration Method for DNNS with Robustness to Domain Shift