Imperfect Verifiers, Exact Targets: Inference-Time Scaling for Continuous Generation

arXiv:2610.15281 · cs.LG, stat.CO, stat.ML · Submitted 2026-10-11 · Read on arXiv

cs.LG, stat.CO, stat.ML

Submitted: 2026-10-11

Updated: 2026-10-11

Code: https://github.com/KraitGit/CS_VGB

Project page: https://liuziwei7.github.io/projects/FaceAttributes.html

Terminology

• Verifier — 1× in this paper · explained in 2 episodes

Sources

• Qwen2.5-VL Technical Report
• Markov operators generated by symmetric measures
• Classifier-Free Diffusion Guidance
• Diamond Maps: Efficient Reward Alignment via Stochastic Flow Maps
• VGB for Masked Diffusion Model: Efficient Test-time Scaling for Reward Satisfaction and Sample Editing
• Effective Test-Time Scaling of Discrete Diffusion through Iterative Refinement
• Strong Stochastic Flow Maps
• Meta Flow Maps enable scalable reward alignment
• Test-time scaling of diffusions with flow maps
• Human Preference Score v2: A Solid Benchmark for Evaluating Human Preferences of Text-to-Image Synthesis
• Training-Free Adaptation of Diffusion Models via Doob's $h$-Transform

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