Comparing SGLD and a fixed-noise predictor-corrector adaptation in canonical joint energy-based models on CIFAR-10

arXiv:2608.05025 · cs.LG, stat.ML · Submitted 2026-08-05 · Read on arXiv

cs.LG, stat.ML

Submitted: 2026-08-05

Updated: 2026-10-14

Comments: 21 pages, 4 figures. v2: revised after peer review; title changed. Under review at Discover Computing

Code: https://github.com/DmytroKnopov/jem-sampler-comparison

License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Terminology

• AUROC — 53× in this paper · explained in 3 episodes
• Accuracy — 18× in this paper · explained in 1 episode
• Failure Modes — 11× in this paper · explained in 3 episodes
• Langevin Dynamics — 6× in this paper · explained in 3 episodes
• Hyperparameters — 5× in this paper · explained in 2 episodes
• Editor — 4× in this paper · explained in 2 episodes
• Joint Energy-Based Models — 4× in this paper
• Theoretical Guarantees — 4× in this paper · explained in 2 episodes
• Adam — 3× in this paper · explained in 3 episodes
• Benchmark — 3× in this paper · explained in 2 episodes
• Diffusion Model — 2× in this paper · explained in 10 episodes
• Fidelity — 2× in this paper · explained in 2 episodes
• Generative Model — 2× in this paper · explained in 2 episodes
• Generative Modeling — 2× in this paper · explained in 3 episodes
• Generative Models — 2× in this paper · explained in 3 episodes
• Out-of-Distribution (OOD) Detection — 2× in this paper · explained in 3 episodes
• Robustness — 2× in this paper · explained in 6 episodes
• Augmentation — 1× in this paper · explained in 2 episodes
• Composition — 1× in this paper · explained in 2 episodes
• Consistency — 1× in this paper · explained in 3 episodes
• Diffusion Models — 1× in this paper · explained in 9 episodes
• Diffusion Process — 1× in this paper · explained in 2 episodes
• Generator — 1× in this paper · explained in 2 episodes
• Performance Gap — 1× in this paper · explained in 2 episodes
• Two-stage training — 1× in this paper · explained in 2 episodes

Sources

• Classifier-Free Guidance is a Predictor-Corrector
• Your Classifier Can Do More: Towards Balancing the Gaps in Classification, Robustness, and Generation
• Scalable Energy-Based Models via Adversarial Training: Unifying Discrimination and Generation

Related papers

• Polynomial-Augmented Neural Networks (PANNs) with Weak Orthogonality Constraints for Enhanced Function and PDE Approximation
• AIRL-S: Unifying Reinforcement Learning and Search-Based Test-Time Scaling via Adversarial Inverse Reinforcement Learning
• Transformers as Bayesian In-Context Experimenters: Smoothness-Adaptive Efficient ATE Estimation
• Convergence issues in Relational Concept Analysis based on AOC-posets
• Beliefs Beyond Posteriors: The Predictive Cost of ELBO Binding
• Understanding Diffusion Models via Ratio-Based Function Approximation with SignReLU Networks