PQSG is a hierarchical question-graph pipeline for fine-grained physical plausibility evaluation in text-to-video generation that achieves higher correlation with human judgments than prior methods on the FinePhyEval dataset.
arXiv preprint arXiv:2506.01943 (2025)
11 Pith papers cite this work. Polarity classification is still indexing.
11
Pith papers citing it
citation-role summary
background 2
citation-polarity summary
years
2026 11roles
background 2polarities
background 2representative citing papers
DSSP is a history-conditioned diffusion state space policy that uses SSMs to encode full observation streams with an auxiliary dynamics objective and hierarchical fusion, achieving SOTA results with reduced model size in robot manipulation.
Mem-World augments world models with W-VMem, a wrist-view-centered surfel memory, to generate persistent action-conditioned video rollouts that improve policy evaluation correlation by 14.5% and raise task success from 58% to 72%.
A single-view video-to-4D pipeline that synthesizes multi-view videos via a T-V attention diffusion model trained on a new Real-MV-4D dataset, then lifts them to 4DGS with flow matching distillation.
A co-evolutionary VLM-VGM loop on 500 unlabeled images raises planner success by 30 points and simulator success by 48 percent while beating fully supervised baselines.
OmniHumanoid factorizes transferable motion learning from embodiment-specific adaptation to enable scalable cross-embodiment video generation without paired data for new humanoids.
SceneScribe-1M is a new dataset of 1 million videos with semantic text, camera parameters, dense depth, and consistent 3D point tracks to support monocular depth estimation, scene reconstruction, point tracking, and text-to-video synthesis.
A dual-tower 4D embodied world model called RoboStereo reduces geometric hallucinations and delivers over 97% relative improvement on manipulation tasks via test-time augmentation, imitative learning, and open exploration.
Visual generation models are evolving from passive renderers to interactive agentic world modelers, but current systems lack spatial reasoning, temporal consistency, and causal understanding, with evaluations overemphasizing perceptual quality.
This survey reviews trends, challenges, benchmarks, and future directions in action-conditioned interactive world modeling for video and 3D generation.
citing papers explorer
-
Physics Question Scene Graph: Fine-grained Evaluation of Physical Plausibility in Text-to-Video Generation
PQSG is a hierarchical question-graph pipeline for fine-grained physical plausibility evaluation in text-to-video generation that achieves higher correlation with human judgments than prior methods on the FinePhyEval dataset.
-
DSSP: Diffusion State Space Policy with Full-History Encoding
DSSP is a history-conditioned diffusion state space policy that uses SSMs to encode full observation streams with an auxiliary dynamics objective and hierarchical fusion, achieving SOTA results with reduced model size in robot manipulation.
-
Mem-World: Memory-Augmented Action-Conditioned World Models for Persistent Robot Manipulation
Mem-World augments world models with W-VMem, a wrist-view-centered surfel memory, to generate persistent action-conditioned video rollouts that improve policy evaluation correlation by 14.5% and raise task success from 58% to 72%.
-
Full-4D: Generating Full-Scope 4D Scenes from a Single-View Video
A single-view video-to-4D pipeline that synthesizes multi-view videos via a T-V attention diffusion model trained on a new Real-MV-4D dataset, then lifts them to 4DGS with flow matching distillation.
-
RoboEvolve: Co-Evolving Planner-Simulator for Robotic Manipulation with Limited Data
A co-evolutionary VLM-VGM loop on 500 unlabeled images raises planner success by 30 points and simulator success by 48 percent while beating fully supervised baselines.
-
OmniHumanoid: Streaming Cross-Embodiment Video Generation with Paired-Free Adaptation
OmniHumanoid factorizes transferable motion learning from embodiment-specific adaptation to enable scalable cross-embodiment video generation without paired data for new humanoids.
-
SceneScribe-1M: A Large-Scale Video Dataset with Comprehensive Geometric and Semantic Annotations
SceneScribe-1M is a new dataset of 1 million videos with semantic text, camera parameters, dense depth, and consistent 3D point tracks to support monocular depth estimation, scene reconstruction, point tracking, and text-to-video synthesis.
-
RoboStereo: Dual-Tower 4D Embodied World Models for Unified Policy Optimization
A dual-tower 4D embodied world model called RoboStereo reduces geometric hallucinations and delivers over 97% relative improvement on manipulation tasks via test-time augmentation, imitative learning, and open exploration.
-
Visual Generation in the New Era: An Evolution from Atomic Mapping to Agentic World Modeling
Visual generation models are evolving from passive renderers to interactive agentic world modelers, but current systems lack spatial reasoning, temporal consistency, and causal understanding, with evaluations overemphasizing perceptual quality.
-
Towards Interactive Video World Modeling: Frontiers, Challenges, Benchmarks, and Future Trends
This survey reviews trends, challenges, benchmarks, and future directions in action-conditioned interactive world modeling for video and 3D generation.
- GEM-4D: Geometry-Enhanced Video World Models for Robot Manipulation