✨WorldCache: Accelerating World Models for Free via Heterogeneous Token Caching

📝 Summary:
WorldCache speeds up slow diffusion-based world models by addressing token heterogeneity and non-uniform dynamics. It uses curvature-guided prediction and chaotic-prioritized skipping. This achieves up to 3.7 times faster inference with 98 percent rollout quality.

🔹 Publication Date: Published on Mar 6

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2603.06331
• PDF: https://arxiv.org/pdf/2603.06331

==================================

For more data science resources:
✓ https://xn--r1a.website/DataScienceT

#WorldModels #DiffusionModels #AI #MachineLearning #Optimization

✨GigaWorld-Policy: An Efficient Action-Centered World--Action Model

📝 Summary:
GigaWorld-Policy is an action-centered World-Action Model that significantly improves robotic policy learning. It decouples visual and motion representations, using dual supervision from action prediction and video generation. This allows for 9x faster inference and 7% higher task success rates c...

🔹 Publication Date: Published on Mar 18

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2603.17240
• PDF: https://arxiv.org/pdf/2603.17240
• Project Page: https://gigaai-research.github.io/GigaWorld-Policy/
• Github: https://github.com/open-gigaai/giga-world-policy

==================================

For more data science resources:
✓ https://xn--r1a.website/DataScienceT

#Robotics #MachineLearning #WorldModels #DeepLearning #PolicyLearning

✨V-JEPA 2: Self-Supervised Video Models Enable Understanding, Prediction and Planning

📝 Summary:
V-JEPA 2 uses self-supervised learning on web videos and minimal robot data. It excels at video understanding, anticipation, Q&A, and zero-shot robotic planning. This approach yields a powerful world model for physical world planning.

🔹 Publication Date: Published on Jun 11, 2025

🔹 Paper Links:
• arXiv Page: https://arxivexplained.com/papers/v-jepa-2-self-supervised-video-models-enable-understanding-prediction-and-planning
• PDF: https://arxiv.org/pdf/2506.09985
• Github: https://github.com/facebookresearch/vjepa2

✨ Datasets citing this paper:
• https://huggingface.co/datasets/ckadirt/vjxla

✨ Spaces citing this paper:
• https://huggingface.co/spaces/vselvarajijay/vjepa2-latent-prediction
• https://huggingface.co/spaces/aavi21458/vjepa2-latent-prediction

==================================

For more data science resources:
✓ https://xn--r1a.website/DataScienceT

#AI #SelfSupervisedLearning #VideoAI #Robotics #WorldModels

✨InCoder-32B-Thinking: Industrial Code World Model for Thinking

📝 Summary:
Industrial software development lacks expert reasoning traces for hardware constraints, so a model was trained on error-driven reasoning chains and domain-specific execution traces to generate high-qu...

🔹 Publication Date: Published on Apr 3

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2604.03144
• PDF: https://arxiv.org/pdf/2604.03144

==================================

For more data science resources:
✓ https://xn--r1a.website/DataScienceT

#AI #CodeGeneration #IndustrialAI #WorldModels #SoftwareDevelopment

✨DriveDreamer-Policy: A Geometry-Grounded World-Action Model for Unified Generation and Planning

📝 Summary:
DriveDreamer-Policy is a unified driving world-action model. It integrates depth, future video, and motion planning using geometry-aware world representation learning. This improves imagined futures and driving actions, achieving strong performance on navigation benchmarks.

🔹 Publication Date: Published on Apr 2

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2604.01765
• PDF: https://arxiv.org/pdf/2604.01765
• Project Page: https://drivedreamer-policy.github.io/
• Github: https://github.com/youngzhou1999/DriveDreamer-Policy

==================================

For more data science resources:
✓ https://xn--r1a.website/DataScienceT

#AutonomousDriving #MotionPlanning #WorldModels #DeepLearning #ComputerVision

✨OpenWorldLib: A Unified Codebase and Definition of Advanced World Models

📝 Summary:
OpenWorldLib presents a standardized framework for advanced world models. It defines a world model as a perception-centered system with interaction and long-term memory for understanding and predicting complex worlds. This unified framework enables efficient model reuse and collaborative inferenc...

🔹 Publication Date: Published on Apr 6

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2604.04707
• PDF: https://arxiv.org/pdf/2604.04707
• Project Page: https://wcny4qa9krto.feishu.cn/wiki/XtPJwf5XQipP7RkeVv0ckyWlnNd

==================================

For more data science resources:
✓ https://xn--r1a.website/DataScienceT

#WorldModels #AI #MachineLearning #DeepLearning #AIFrameworks

✨dWorldEval: Scalable Robotic Policy Evaluation via Discrete Diffusion World Model

📝 Summary:
dWorldEval proposes a scalable robotics policy evaluation method using a discrete diffusion world model. It unifies diverse modalities into a token space, employing a transformer and progress token for success detection. This approach significantly outperforms prior methods, enabling large-scale ...

🔹 Publication Date: Published on Apr 24

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2604.22152
• PDF: https://arxiv.org/pdf/2604.22152

==================================

For more data science resources:
✓ https://xn--r1a.website/DataScienceT

#Robotics #DiffusionModels #WorldModels #AI #MachineLearning

🔥 LeWorldModel: Stable End-to-End Joint-Embedding Predictive Architecture from Pixels

💡 The paper introduces LeWorldModel, a stable end to end joint embedding predictive architecture that trains efficiently from raw pixels. Existing methods for learning world models in compact latent spaces are fragile and rely on complex loss terms, pre trained encoders, or auxiliary supervision to avoid representation collapse. LeWorldModel addresses this issue by using only two loss terms, a next embedding prediction loss and a regularizer, to train the model end to end from raw pixels. This approach reduces the number of tunable loss hyperparameters from six to one compared to existing methods. The model has approximately 15 million parameters and can be trained on a single GPU in a few hours, making it up to 48 times faster than foundation model based world models. The results show that LeWorldModel remains competitive across diverse 2D and 3D control tasks and encodes meaningful physical structures in its latent space. The model is also able to reliably detect physically implausible events, demonstrating its ability to learn a robust and generalizable representation of the world. Overall, LeWorldModel provides a stable and efficient framework for learning world models from raw pixels, making it a significant contribution to the field of artificial intelligence.

📅 Published on Mar 13

🔗 Links:
• arXiv: https://arxiv.org/abs/2603.19312
• PDF: https://arxiv.org/pdf/2603.19312
• Project Page: https://le-wm.github.io/
• GitHub: https://github.com/lucas-maes/le-wm ⭐ 3.1k

🤖 Models citing this paper:
• https://huggingface.co/quentinll/lewm-pusht
• https://huggingface.co/aguennoune17/atlas-v2-nwm-fp8-compressed
• https://huggingface.co/quentinll/lewm-tworooms

📊 Datasets citing this paper:
• https://huggingface.co/datasets/quentinll/lewm-pusht
• https://huggingface.co/datasets/quentinll/lewm-cube
• https://huggingface.co/datasets/quentinll/lewm-reacher

━━━━━━━━━━━━━━━━━━━━━━━━
📢 By: https://xn--r1a.website/PaperNexus

#WorldModels #JointEmbedding #PredictiveArchitectures #EndToEndLearning #LatentSpaceRepresentation

🔥 World Model for Robot Learning: A Comprehensive Survey

💡 The paper provides a comprehensive survey of world models for robot learning, which are predictive representations of environmental dynamics that support policy learning, planning, and simulation. The authors note that the literature on world models is fragmented across different architectures, functional roles, and application domains, making it difficult to understand the current state of the field. To address this gap, the authors present a systematic review of world models from a robot learning perspective, examining how they are coupled with robot policies, used as learned simulators for reinforcement learning and evaluation, and have progressed in terms of robotic video world models. The survey covers the progression of world models from imagination-based generation to controllable, structured, and foundation-scale formulations, and connects these ideas to navigation and autonomous driving. The authors also summarize representative datasets, benchmarks, and evaluation protocols, and highlight major challenges and future directions for predictive modeling in embodied agents. The paper aims to clarify key paradigms and applications of world models, and to facilitate continued access to newly emerging works, benchmarks, and resources, the authors will maintain and regularly update a GitHub repository accompanying the survey. Overall, the paper provides a thorough overview of the rapidly growing literature on world models for robot learning, and identifies key areas for future research and development.

📅 Published on Apr 30

🔗 Links:
• arXiv: https://arxiv.org/abs/2605.00080
• PDF: https://arxiv.org/pdf/2605.00080
• Project Page: https://ntumars.github.io/wm-robot-survey/
• GitHub: https://github.com/NTUMARS/Awesome-World-Model-for-Robotics-Policy ⭐ 317

━━━━━━━━━━━━━━━━━━━━━━━━
📢 By: https://xn--r1a.website/PaperNexus

#RobotLearning #WorldModels #PredictiveRepresentations #ReinforcementLearning #RobotPolicies

🔥 Cosmos 3: Omnimodal World Models for Physical AI

💡 The paper introduces Cosmos 3, an omnimodal world model that can process and generate multiple data types, including language, image, video, audio, and action sequences, through a unified mixture-of-transformers architecture. This model is designed to jointly process and generate different modalities, effectively combining vision-language models, video generators, world simulators, and world-action models into a single framework. The Cosmos 3 model achieves state-of-the-art performance in various understanding and generation tasks, demonstrating its ability to serve as a scalable and general-purpose backbone for embodied agents.

The key contribution of the paper is the development of a unified architecture that can handle multiple input-output configurations, allowing for seamless integration of different modalities. The model is evaluated on a diverse suite of tasks, including text-to-image and image-to-video generation, and policy modeling, and achieves superior performance compared to existing models. The authors also make their code, model checkpoints, curated synthetic datasets, and evaluation benchmark available under an open-source license to accelerate research and deployment in Physical AI.

The results show that the Cosmos 3 model establishes a new state-of-the-art in various tasks, and its post-trained models were ranked as the best open-source models in text-to-image and image-to-video generation, as well as policy modeling. The availability of the code and model checkpoints is expected to facilitate further research and development in Physical AI, and the Cosmos 3 model has the potential to become a widely-used backbone for embodied agents. Overall, the paper presents a significant contribution to the field of Physical AI, demonstrating the effectiveness of omnimodal world models in achieving state-of-the-art performance in a wide range of tasks.

📅 Published on Jun 1

🔗 Links:
• GitHub: https://github.com/huggingface
• arXiv: https://arxiv.org/abs/2606.02800
• PDF: https://arxiv.org/pdf/2606.02800
• Project Page: https://research.nvidia.com/labs/cosmos-lab/cosmos3/

━━━━━━━━━━━━━━━━━━━━━━━━
📢 By: https://xn--r1a.website/PaperNexus

#OmnimodalLearning #MultimodalGeneration #PhysicalAI #WorldModels #EmbodiedIntelligence

🔥 INTACT: Isomorphic Intent-to-Action Learning for Search-Free World Models

💡 The paper introduces INTACT, a novel approach to learning world models that can predict how actions change a scene and recover actions for a desired change without expensive test-time search. The existing forward latent world models predict how actions change a scene, but recovering actions for a desired change is only possible through expensive test-time search. INTACT turns action-labeled, reward-free trajectories into a deployable intent-to-action interface. Each transition supplies physical intent, while a future goal supplies deployment intent. The architecture is isomorphic between local and goal motion-intent backbone-input graphs through an identical four-slot grammar and shared parameters.

The method provides intact transfer from RGB evidence to action-effective latent intent coordinates and from intent families to their corresponding action-law families. Asymmetric endpoint gradients ground physical successors and fix future goals as anchors, joining representation learning and control without pointwise latent matching or globally linear dynamics. The resulting coordinates support a robust distributional action law, where its conditional means serve directly as a search-free policy, while sampling remains available for diversity or optional verification.

The results show that one-epoch, zero-search models reach high success rates on four official tasks, achieving 85.78, 100.00, 97.67, and 97.89 percent success. Optional local cross-entropy method centered on the Direct plan reaches 96.86 percent macro success using 384 instead of 9000 candidate sequences, reducing sampling by 23.44 times while improving pure cross-entropy method by 16.00 points. One shared four-task encoder reaches 89.39 percent E5 Direct macro and improves every task over jointly trained models, while predicted expert action-family kNN tracks Direct success at 0.954. Direct inference takes 2.9-5.5 milliseconds. Overall, INTACT provides a robust and efficient approach to learning world models that can predict and recover actions for desired changes without expensive test-time search.

📅 Published on Jul 28

🔗 Links:
• GitHub: https://github.com/huggingface
• arXiv: https://arxiv.org/abs/2607.26056
• PDF: https://arxiv.org/pdf/2607.26056
• Project Page: https://zju3dv.github.io/INTACT-JEPA/

━━━━━━━━━━━━━━━━━━━━━━━━
📢 By: https://xn--r1a.website/PaperNexus

#IntentToActionLearning #WorldModels #IsomorphicLearning #SearchFreePlanning #LatentWorldModels

🔥 PlayWorld: Benchmarking World Models with Agent Players over Long-Horizon Objectives

💡 The paper introduces PlayWorld, a benchmark for evaluating interactive video world models. The problem addressed is that current methods for comparing these models are inadequate, as they typically rely on fixed action-conditioned evaluation, which is not suitable for cross-model comparison. To address this, the authors propose using multi-modal agents, called Agent Players, to interact with world models towards long-horizon objectives, such as turning around 360 degrees to check environment consistency or walking into water to see realistic water ripples.

The method involves employing these Agent Players to pursue specified objectives in 171 scenarios, and evaluating the models along four core dimensions: geometry consistency, interaction fidelity, out-of-sight evolution, and insight evolution. Additionally, basic ability metrics for video quality and controllability are assessed.

The results of the experiments, which were conducted on nine state-of-the-art world models, show that current models are unreliable on long-horizon interactive objectives, particularly in maintaining spatial consistency and persistent state evolution. The paper contributes a new benchmark, PlayWorld, which provides a comprehensive evaluation framework for interactive video world models, and highlights the need for further research in this area. The code and data are made available for further use.

📅 Published on Aug 13

🔗 Links:
• GitHub: https://github.com/huggingface
• arXiv: https://arxiv.org/abs/2608.13552
• PDF: https://arxiv.org/pdf/2608.13552
• Project Page: https://kxding.github.io/project/PlayWorld/

📊 Datasets citing this paper:
• https://huggingface.co/datasets/jocelynd/playworld-bench

🚀 Spaces citing this paper:
• https://huggingface.co/spaces/jocelynd/PlayWorld-Leaderboard

━━━━━━━━━━━━━━━━━━━━━━━━
📢 By: https://xn--r1a.website/PaperNexus

#WorldModels #LongHorizonObjectives #AgentPlayers #InteractiveVideoModels #BenchmarkingWorldModels