PAWBench: How Far Are We from Probabilistically Aligned World Modeling?

arXiv:2608.27345 · cs.CV, cs.AI · Submitted 2026-08-27 · Read on arXiv

cs.CV, cs.AI

Submitted: 2026-08-27

Updated: 2026-09-03

Code: https://github.com/Wan-Video/Wan2.2

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

Terminology

• Calibration — 54× in this paper · explained in 7 episodes
• Benchmark — 27× in this paper · explained in 2 episodes
• Alignment — 26× in this paper · explained in 1 episode
• World Model — 26× in this paper · explained in 7 episodes
• Aggregation — 7× in this paper · explained in 1 episode
• Generative Model — 6× in this paper · explained in 2 episodes
• Generative Models — 6× in this paper · explained in 2 episodes
• Adaptation — 4× in this paper · explained in 1 episode
• Benchmarking — 4× in this paper · explained in 3 episodes
• Consistency — 3× in this paper · explained in 3 episodes
• Fine-Tuning — 3× in this paper · explained in 5 episodes
• Accuracy — 2× in this paper · explained in 1 episode
• Computer Vision and Pattern Recognition — 2× in this paper
• Fidelity — 2× in this paper · explained in 2 episodes
• Foundation Model — 2× in this paper · explained in 3 episodes
• Inference-Time Steering — 2× in this paper · explained in 2 episodes
• Prompt Engineering — 2× in this paper · explained in 4 episodes
• Scene Pass Rate — 2× in this paper
• Steering — 2× in this paper · explained in 2 episodes
• Alignment Scores — 1× in this paper · explained in 1 episode
• Failure Modes — 1× in this paper · explained in 3 episodes
• LoRA Adaptation — 1× in this paper · explained in 2 episodes
• Monte Carlo Analysis — 1× in this paper · explained in 2 episodes
• Reinforcement Learning — 1× in this paper · explained in 4 episodes
• Robustness — 1× in this paper · explained in 5 episodes
• Stochastic Dynamics — 1× in this paper · explained in 2 episodes
• Temporal consistency — 1× in this paper · explained in 2 episodes

Sources

• Cosmos 3: Omnimodal World Models for Physical AI
• VideoPhy: Evaluating Physical Commonsense for Video Generation
• Genie: Generative Interactive Environments
• On Calibration of Modern Neural Networks
• T2VPhysBench: A First-Principles Benchmark for Physical Consistency in Text-to-Video Generation
• World Models
• LTX-2: Efficient Joint Audio-Visual Foundation Model
• VBench++: Comprehensive and Versatile Benchmark Suite for Video Generative Models
• Couple to Control: Joint Initial Noise Design in Diffusion Models
• How Far is Video Generation from World Model: A Physical Law Perspective
• Improved Precision and Recall Metric for Assessing Generative Models
• Scaling Mixture-of-Experts Video Pretraining for Embodied Intelligence
• How Confident are Video Models? Empowering Video Models to Express their Uncertainty
• World Models That Know When They Don't Know - Controllable Video Generation with Calibrated Uncertainty
• Towards World Simulator: Crafting Physical Commonsense-Based Benchmark for Video Generation
• String Seed of Thought: Prompting LLMs for Distribution-Faithful and Diverse Generation
• Reliable Fidelity and Diversity Metrics for Generative Models
• Do Deep Generative Models Know What They Don't Know?
• Cosmos World Foundation Model Platform for Physical AI
• PICABench: How Far Are We from Physically Realistic Image Editing?

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