PreviewDiff: Multimodal Critic-Guided Search over Diffusion Latents

arXiv:2609.36199 · cs.CV, cs.AI, cs.LG · Submitted 2026-09-28 · Read on arXiv

cs.CV, cs.AI, cs.LG

Submitted: 2026-09-28

Updated: 2026-09-28

Project page: https://previewdiff.github.io

Terminology

• Verifier — 22× in this paper · explained in 2 episodes
• Benchmark — 21× in this paper · explained in 2 episodes
• Diffusion Model — 15× in this paper · explained in 8 episodes
• Ablation — 12× in this paper · explained in 1 episode
• Diffusion Models — 12× in this paper · explained in 6 episodes
• Test-time Scaling — 12× in this paper · explained in 2 episodes
• Alignment — 8× in this paper · explained in 1 episode
• Decode — 8× in this paper · explained in 2 episodes
• Prompt Adherence — 5× in this paper · explained in 2 episodes
• Steering — 5× in this paper · explained in 2 episodes
• Pruning — 4× in this paper · explained in 4 episodes
• Action Space — 2× in this paper · explained in 1 episode
• Consistency — 2× in this paper · explained in 3 episodes
• Best-of-N Sampling — 1× in this paper · explained in 2 episodes
• Decoupling — 1× in this paper · explained in 2 episodes
• Failure Modes — 1× in this paper · explained in 3 episodes
• Fidelity — 1× in this paper · explained in 2 episodes
• Generative Model — 1× in this paper · explained in 2 episodes
• Generative Models — 1× in this paper · explained in 2 episodes
• Grounding — 1× in this paper · explained in 3 episodes
• Robustness — 1× in this paper · explained in 5 episodes
• Temporal consistency — 1× in this paper · explained in 2 episodes
• Training-Free Framework — 1× in this paper · explained in 2 episodes
• Verification — 1× in this paper · explained in 2 episodes

Sources

• Qwen3-VL Technical Report
• NarrLV: Towards a Comprehensive Narrative-Centric Evaluation for Long Video Generation
• LTX-Video: Realtime Video Latent Diffusion
• Anchored Video Generation: Decoupling Scene Construction and Temporal Synthesis in Text-to-Video Diffusion Models
• Scaling Image and Video Generation via Test-Time Evolutionary Search
• Classifier-Free Diffusion Guidance
• Diffusion Tree Sampling: Scalable inference-time alignment of diffusion models
• GenEval 2: Addressing Benchmark Drift in Text-to-Image Evaluation
• Test-time Alignment of Diffusion Models without Reward Over-optimization
• Easier Painting Than Thinking: Can Text-to-Image Models Set the Stage, but Not Direct the Play?
• Inference-Time Scaling for Diffusion Models beyond Scaling Denoising Steps
• Hierarchical Text-Conditional Image Generation with CLIP Latents
• Test-Time Scaling of Diffusion Models via Noise Trajectory Search
• A General Framework for Inference-time Scaling and Steering of Diffusion Models
• Scaling LLM Test-Time Compute Optimally can be More Effective than Scaling Model Parameters
• UniGen: Enhanced Training & Test-Time Strategies for Unified Multimodal Understanding and Generation
• Wan: Open and Advanced Large-Scale Video Generative Models
• Monte Carlo Tree Diffusion for System 2 Planning
• VBench-2.0: Advancing Video Generation Benchmark Suite for Intrinsic Faithfulness

Related papers

• Loss Knows Best: Detecting Annotation Errors in Videos via Loss Trajectories
• AnchorWeave: World-Consistent Video Generation with Retrieved Local Spatial Memories
• Benchmarking the Robustness of Foundation Models for Mammography under Domain Shift
• MambaX-Net: Dual-Input Mamba-Enhanced Cross-Attention Network for Longitudinal MRI Segmentation
• TeleOCR: Navigating Document Parsing Across Digital and Camera-Captured Documents
• A Survey on Efficient Vision-Language-Action Models