arXiv:2608. 05948v1 Announce Type: new Abstract: Physics engines facilitate large-scale training and evaluation for embodied intelligence, while generative video world models are emerging as implicit simulators of future states and interactions.
Physics engines facilitate large-scale training and evaluation for embodied intelligence, while generative video world models are emerging as implicit simulators of future states and interactions. However, existing evaluations of physical fidelity are often conducted in isolation and rely heavily on perceptual similarity or human judgments, providing limited insight into which physical principles or parameters are violated.
arXiv:2609.24308v1 Announce Type: new Abstract: Evaluating world models requires assessing both the quality of the worlds they generate and their consistency and responsiveness under exploration, int...
Embodied intelligence and world models require video understanding systems to go beyond recognizing objects and actions and develop an understanding of physical regularities. However, despite their strong performance on general video understanding tasks, current video-language models still struggle to reliably determine whether an observed event conforms to specific physical laws.
arXiv:2608. 02150v2 Announce Type: replace-cross Abstract: Embodied intelligence and world models require video understanding systems to go beyond recognizing objects and actions and develop an understanding of physical regularities.
PhysicsLENS is a new dataset and benchmark designed to evaluate how well video generation models capture physical properties relevant to robotics. It consists of matched scenario pairs that share the same conditioning frame and task but differ in underlying physics, and it covers seven physical domains such as collision, gravity, and friction. The benchmark includes over 400 human-annotated labels from four video generation models, revealing that many plausible-looking videos still ignore the specified physical property.
The World Embedding Benchmark introduces 8,000 simulation-based video cases covering fluid, solid, dynamic, and optical physics, each paired with physical annotations. It supports three tasks—text‑video retrieval, physical‑property regression, and multiple‑choice classification—to assess how well video embeddings capture physical alignment versus quantitative information. Experiments show that while pre‑trained models perform poorly on retrieval and classification, lightweight probes can extract useful physical data, and physics‑specific contrastive training improves alignment but harms regression, highlighting a trade‑off. Retrieval‑augmented generation using these embeddings further enhances the physical fidelity of generated videos.
RoboPhys-3D is a 3D‑grounded embodied world model benchmark built on RoboTwin 2.0, featuring 50 manipulation tasks, 5,000 episodes, and 25,000 multi‑view ground‑truth videos. It evaluates video world models by processing both generated and ground‑truth videos through the same 3D reconstruction pipeline, allowing the separation of reconstruction‑induced from generation‑induced errors. The benchmark defines 50 metrics across four sub‑dimensions—pixel fidelity, 3D geometry consistency, state understanding, and task completeness—and introduces the Average Full Score and RoboPhyscore for holistic assessment, with RoboPhyscore showing strong correlation with human judgments.
Evaluating world models requires assessing both the quality of the worlds they generate and their consistency and responsiveness under exploration, interaction, and modification. We introduce HappyWor...
arXiv:2606. 28128v1 Announce Type: cross Abstract: Video generation models have emerged as a promising paradigm for embodied world simulation.
Ego2Act is a new benchmark that tests video generation models on goal‑directed, egocentric manipulation tasks. It contains 2,640 videos from 110 real‑world tasks, each requiring multiple steps of object manipulation to achieve a high‑level goal. The benchmark includes Ego2ActJudge, a reference‑free evaluation pipeline that better aligns with human judgments of task completion and physics plausibility.
arXiv:2608. 09298v1 Announce Type: cross Abstract: Action-conditioned world models (ACWMs) promise to provide embodied AI with scalable predictive simulators for planning, policy evaluation, and data generation.