Research map / Image generation / Autoregressive and latent generation
Feature Likelihood Divergence: Evaluating the Generalization of Generative Models Using Samples
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Replace generated samples with imperceptibly transformed copies | PFGM++ samples evaluated with FLD · Image transformations | FLD | Mixed |
| Apply large structural transformations such as cropping or rotation | PFGM++ samples evaluated with FLD · Image transformations | FLD | Hurt |
| Reduce the number of represented classes or increase duplicate samples | Conditional generative model samples · CIFAR10 | FLD | Hurt |
| Replace generated samples with transformed copies of training data | Generative model samples evaluated with FLD · CIFAR10 | Generalization gap FLD | Hurt |
| Use DINOv2 rather than Inception-V3 features | FLD and other sample-based metrics · Image benchmarks | Correlation with human judgment | Helped |
Closest papers by meaning
- Assessing Generative Models via Precision and Recall (NeurIPS 2018 · similarity 0.91)
- Image Generation Diversity Issues and How to Tame Them (CVPR 2025 · similarity 0.88)
- Direct Discriminative Optimization: Your Likelihood-Based Visual Generative Model is Secretly a GAN Discriminator (ICML 2025 · similarity 0.87)
- Forte : Finding Outliers with Representation Typicality Estimation (ICLR 2025 · similarity 0.87)
- Unveiling Differences in Generative Models: A Scalable Differential Clustering Approach (CVPR 2025 · similarity 0.87)
- Attribute Based Interpretable Evaluation Metrics for Generative Models (ICML 2024 · similarity 0.87)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.