mlsanityData to 7 Oct 2026

DriftLite: Lightweight Drift Control for Inference-Time Scaling of Diffusion Models

arXiv 2509.21655 ↗First posted 25 Sep 2025cs.LG stat.MLcited by 21

On the map: LLM reasoning → Diffusion reasoning and test-time scaling

also in 1 more
Abstract (arXiv)

We study inference-time scaling for diffusion models, where the goal is to adapt a pre-trained model to new target distributions without retraining. Existing guidance-based methods are simple but introduce bias, while particle-based corrections suffer from weight degeneracy and high computational cost. We introduce DriftLite, a lightweight, training-free particle-based approach that steers the inference dynamics on the fly with provably optimal stability control. DriftLite exploits a previously unexplored degree of freedom in the Fokker-Planck equation between the drift and particle potential, and yields two practical instantiations: Variance- and Energy-Controlling Guidance (VCG/ECG) for approximating the optimal drift with minimal overhead. Across Gaussian mixture models, particle systems, and large-scale protein-ligand co-folding problems, DriftLite consistently reduces variance and improves sample quality over pure guidance and sequential Monte Carlo baselines. These results highlight a principled, efficient route toward scalable inference-time adaptation of diffusion models. Our source code is publicly available at https://github.com/yinuoren/DriftLite.

Key facts. Cited by 21 papers in the map (data through 7 Oct 2026). Placed in Diffusion reasoning and test-time scaling, one of 17 directions in LLM reasoning. Builds on 54 of its 117 references that are in the map. First posted on arXiv 25 Sep 2025.

Builds on

54 of 117 references.

Closest ancestors · 2

Same direction, newest first

  1. 2025Inference-Time Scaling for Diffusion Models beyond Scaling Denoising Steps16 Jan
  2. 2025A General Framework for Inference-time Scaling and Steering of Diffusion Models12 Jan

Other references · 52

  1. 2014Adam: A Method for Stochastic Optimizationcited by 36,960
  2. 2020Score-Based Generative Modeling through Stochastic Differential Equationscited by 8,964
  3. 2022Classifier-Free Diffusion Guidancecited by 5,717
  4. 2022Flow Matching for Generative Modelingcited by 5,426
  5. 2021GLIDE: Towards Photorealistic Image Generation and Editing with Text-Guided Diffusion Modelscited by 3,595
  6. 2022Flow Straight and Fast: Learning to Generate and Transfer Data with Rectified Flowcited by 3,550
  7. 2022Diffusion Posterior Sampling for General Noisy Inverse Problemscited by 1,375
  8. 2022Building Normalizing Flows with Stochastic Interpolantscited by 964
  9. 2023Training Diffusion Models with Reinforcement Learningcited by 939
  10. 2023Discrete Diffusion Modeling by Estimating the Ratios of the Data Distributioncited by 712
  11. 2022GeoDiff: a Geometric Diffusion Model for Molecular Conformation Generationcited by 511
  12. 2023Directly Fine-Tuning Diffusion Models on Differentiable Rewardscited by 431
38 more
  1. 2022Diffusion probabilistic modeling of protein backbones in 3D for the motif-scaffolding problemcited by 229
  2. 2024Adjoint Matching: Fine-tuning Flow and Diffusion Generative Models with Memoryless Stochastic Optimal Controlcited by 222
  3. 2021Path Integral Sampler: a stochastic control approach for samplingcited by 175
  4. 2023Manifold Preserving Guided Diffusioncited by 153
  5. 2023Denoising Diffusion Samplerscited by 144
  6. 2024Fine-Tuning of Continuous-Time Diffusion Models as Entropy-Regularized Controlcited by 136
  7. 2024Iterated Denoising Energy Matching for Sampling from Boltzmann Densitiescited by 113
  8. 2023Improved sampling via learned diffusionscited by 103
  9. 2024Probabilistic Inference in Language Models via Twisted Sequential Monte Carlocited by 85
  10. 2025Feynman-Kac Correctors in Diffusion: Annealing, Guidance, and Product of Expertscited by 85
  11. 2023Optimizing DDPM Sampling with Shortcut Fine-Tuningcited by 81
  12. 2024Provably Robust Score-Based Diffusion Posterior Sampling for Plug-and-Play Image Reconstructioncited by 73
  13. 2024NETS: A Non-Equilibrium Transport Samplercited by 67
  14. 2025InverseBench: Benchmarking Plug-and-Play Diffusion Priors for Inverse Problems in Physical Sciencescited by 67
  15. 2024Gradient Guidance for Diffusion Models: An Optimization Perspectivecited by 59
  16. 2024How Discrete and Continuous Diffusion Meet: Comprehensive Analysis of Discrete Diffusion Models via a Stochastic Integral Frameworkcited by 48
  17. 2025Fast Solvers for Discrete Diffusion Models: Theory and Applications of High-Order Algorithmscited by 47
  18. 2024Theoretical Insights for Diffusion Guidance: A Case Study for Gaussian Mixture Modelscited by 46
  19. 2024What does guidance do? A fine-grained analysis in a simple settingcited by 45
  20. 2018Monge-Amp`ere Flow for Generative Modelingcited by 33
  21. 2025Debiasing Guidance for Discrete Diffusion with Sequential Monte Carlocited by 29
  22. 2025Progressive Inference-Time Annealing of Diffusion Models for Sampling from Boltzmann Densitiescited by 26
  23. 2023Transport meets Variational Inference: Controlled Monte Carlo Diffusionscited by 21
  24. 2025Adjoint Schr"odinger Bridge Samplercited by 20
  25. 2025MDNS: Masked Diffusion Neural Sampler via Stochastic Optimal Controlcited by 20
  26. 2024Diffusion Models as Cartoonists: The Curious Case of High Density Regionscited by 19
  27. 2024Ensemble Kalman Diffusion Guidance: A Derivative-free Method for Inverse Problemscited by 18
  28. 2024Generative Artificial Intelligence for Navigating Synthesizable Chemical Spacecited by 18
  29. 2025Composition and Control with Distilled Energy Diffusion Models and Sequential Monte Carlocited by 18
  30. 2025Solving Inverse Problems via Diffusion-Based Priors: An Approximation-Free Ensemble Sampling Approachcited by 18
  31. 2025Learning normalized image densities via dual score matchingcited by 17
  32. 2025RNE: plug-and-play diffusion inference-time control and energy-based trainingcited by 16
  33. 2024Posterior Sampling with Denoising Oracles via Tilted Transportcited by 12
  34. 2025Improving Classifier-Free Guidance in Masked Diffusion: Low-Dim Theoretical Insights with High-Dim Impactcited by 11
  35. 2025A Unified Approach to Analysis and Design of Denoising Markov Modelscited by 10
  36. 2024Quantum State Generation with Structure-Preserving Diffusion Modelcited by 8
  37. 2024Guidance for twisted particle filter: a continuous-time perspectivecited by 7
  38. 2025Efficient and Unbiased Sampling from Boltzmann Distributions via Variance-Tuned Diffusion Modelscited by 6

50 most cited of 52 All 52 (JSON)

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2025: 1; 2026: 20.

21 papers.

Sort:

By area:

  1. 2025Enhancing Reasoning for Diffusion LLMs via Distribution Matching Policy Optimizationcited by 15
  2. 2026Enhanced Diffusion Sampling: Efficient Rare Event Sampling and Free Energy Calculation with Diffusion Modelscited by 11
  3. 2026Robust Inference-Time Steering of Protein Diffusion Models via Embedding Optimizationcited by 6
  4. 2026How to Guide Your Flow: Few-Step Alignment via Flow Map Reward Guidancecited by 5
  5. 2026Adjoint Matching through the Lens of the Stochastic Maximum Principle in Optimal Controlcited by 4
  6. 2026Contrastive Distribution Matching for Amortized Sequential Monte Carlo in Discrete Diffusioncited by 3
  7. 2026Inference-Time Alignment for Diffusion Models via Variationally Stable Doob's Matchingcited by 2
  8. 2026Compression as Adaptation: Implicit Visual Representation with Diffusion Foundation Modelscited by 2
  9. 2026Refining Compositional Diffusion for Reliable Long-Horizon Planningcited by 2
  10. 2026Simple Approximation and Derivative Free Inference-Time Scaling for Diffusion Models via Sequential Monte Carlo on Path Measurescited by 2
  11. 2026SURGE: Approximation and Training Free Particle Filter for Diffusion Surrogatecited by 2
  12. 2026Generative Modeling in Protein Design: Neural Representations, Conditional Generation, and Evaluation Standardscited by 1
9 more
  1. 2026Diffusion-Based Posterior Sampling: A Feynman-Kac Analysis of Bias and Stabilitycited by 1
  2. 2026NoiseRater: Meta-Learned Noise Valuation for Diffusion Model Trainingcited by 1
  3. 2026Catastrophic Compositional Generation: Why Vanilla Diffusion Models Fail to Extrapolatecited by 1
  4. 2026FluxLite: Inference-Time Proposal Control for Discrete Diffusion Modelscited by 1
  5. 2026A Mean-Field Framework for Inference-Time Distributional Control of Diffusion Modelscited by 0
  6. 2026METALICA: METAdynamics and repLICA exchange for enhanced diffusion samplingcited by 0
  7. 2026Interacting particle guidance for sampling reward-tilted generative priorscited by 0
  8. 2026Specificity-Aware Diffusion Steering via Variance-Reduced Sequential Monte Carlocited by 0
  9. 2026Error-Corrected Inference-Time Scaling for Imperfect Diffusion Modelscited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2025Inference-Time Alignment in Diffusion Models with Reward-Guided Generation: Tutorial and Review
  2. 2025Tortoise and Hare Guidance: Accelerating Diffusion Model Inference with Multirate Integration
  3. 2026Diversify Diffusion with Temperature Sampling and Variance-Corrective Time Shifting
  4. 2025Navigating the Exploration-Exploitation Tradeoff in Inference-Time Scaling of Diffusion Models
  5. 2025Diffusion Models are Molecular Dynamics Simulators
  6. 2025Inference-Time Scaling of Discrete Diffusion Models via Importance Weighting and Optimal Proposal Design
  7. 2026Stochastic Transition-Map Distillation for Fast Probabilistic Inference
  8. 2025Consistent Sampling and Simulation: Molecular Dynamics with Energy-Based Diffusion Models