Research map / Probabilistic and scientific modeling / Variational and generative sampling
Implicit Generation and Modeling with Energy Based Models
Yilun Du, Igor Mordatch
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 |
|---|---|---|---|
| ensemble sampling from last 10 model snapshots (ablation) | EBM · CIFAR-10 Unconditional | FID: 40.58 -> 38.2 | Improved |
| ensemble sampling from last 10 model snapshots (ablation) | EBM · CIFAR-10 Unconditional | Inception score: 6.02 -> 6.78 | Improved |
| EBM instead of PixelIQN | likelihood generative models · CIFAR-10 Unconditional | FID: 49.46 -> 40.58 | Improved |
| EBM instead of SNGAN | GAN generator · CIFAR-10 Unconditional | FID: 21.7 vs 40.58 | Hurt |
| EBM instead of SNGAN at 128x128 (smaller net, not converged) | ImageNet 128x128 generator · ImageNet 128x128 Conditional | FID: 27.62 vs 43.7 | Hurt |
| EBM instead of PixelIQN | conditional generative models · ImageNet 32x32 Conditional | FID: 22.99 -> 14.31 | Improved |
| EBM instead of feedforward FC network | trajectory model · Robot hand, unconditional | average Frechet distance: 33.28 -> 5.96 | Improved |
| EBM instead of feedforward FC network | trajectory model · Robot hand, conditional | average Frechet distance: 19.75 -> 8.97 | Improved |
| EBM instead of LwF | continual learning · Split MNIST | accuracy: 24.17 -> 64.99 | Improved |
| EBM instead of conditional VAE with same likelihood-classification scheme | generative continual learner · Split MNIST | accuracy: 40.04 -> 64.99 | Improved |
| EBM instead of PixelCNN++ | likelihood models · OOD detection average | AUROC: 0.47 -> 0.62 | Improved |
| run PGD attack incorporating sampling | EBM classifier · CIFAR-10 adversarial robustness | robustness | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- PixelCNN — PixelCNN++: Improving the PixelCNN with Discretized Logistic Mixture Likelihood and Other Modifications (ICLR 2017) · better on 5 of 5 shared results
- Glow — Glow: Generative Flow with Invertible 1x1 Convolutions (NeurIPS 2018) · better on 3 of 3 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Learning Energy-Based Models with Adversarial Training (ECCV 2022 · similarity 0.90)
- Generalized Energy Based Models (ICLR 2021 · similarity 0.89)
- Scalable Energy-Based Models via Adversarial Training: Unifying Discrimination and Generation (ICLR 2026 · similarity 0.88)
- Learning Energy-Based Models by Cooperative Diffusion Recovery Likelihood (ICLR 2024 · similarity 0.87)
- Improving Adversarial Energy-Based Model via Diffusion Process (ICML 2024 · similarity 0.87)
- Energy Matching: Unifying Flow Matching and Energy-Based Models for Generative Modeling (NeurIPS 2025 · similarity 0.86)
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.
