Random Network Distillation as a Diversity Metric for Both Image and Text Generation
arXiv:2010.06715
Abstract
Generative models are increasingly able to produce remarkably high quality images and text. The community has developed numerous evaluation metrics for comparing generative models. However, these metrics do not effectively quantify data diversity. We develop a new diversity metric that can readily be applied to data, both synthetic and natural, of any type. Our method employs random network distillation, a technique introduced in reinforcement learning. We validate and deploy this metric on both images and text. We further explore diversity in few-shot image generation, a setting which was previously difficult to evaluate.
References in corpus (20)
- Adam: A Method for Stochastic Optimization
- PyTorch: An Imperative Style, High-Performance Deep Learning Library
- GANs Trained by a Two Time-Scale Update Rule Converge to a Local Nash Equilibrium
- Large Scale GAN Training for High Fidelity Natural Image Synthesis
- Language Models are Few-Shot Learners
- Self-Attention Generative Adversarial Networks
- Conditional Image Synthesis With Auxiliary Classifier GANs
- Progressive Growing of GANs for Improved Quality, Stability, and Variation
- Pointer Sentinel Mixture Models
- GAN Augmentation: Augmenting Training Data using Generative Adversarial Networks
- Defense-GAN: Protecting Classifiers Against Adversarial Attacks Using Generative Models
- Exploration by Random Network Distillation
- The generalization error of random features regression: Precise asymptotics and double descent curve
- A Note on the Inception Score
- Generalization Properties of Learning with Random Features
- Classification Accuracy Score for Conditional Generative Models
- FIGR: Few-shot Image Generation with Reptile
- New Alignment Methods for Discriminative Book Summarization
- Real or Not Real, that is the Question
- Jointly Measuring Diversity and Quality in Text Generation Models