Skip to content
#Data and simulation
Oct 28Wed
  1. APPSO · 微信公众号71

    Google DeepMind releases world model Genie 4, extending single simulation to 120 seconds

    Google DeepMind has released Genie 4, a new-generation world model that extends single interactive simulation duration from about one minute to two minutes and supports native audio generation for the first time. Genie 4 introduces Gemini to plan actions while it generates the environment in real time; in a demo, Gemini 4 Argon breaks a task into consecutive steps, and Genie 4 simulates the Apptronik Apollo humanoid robot playing checkers.

  2. 机器人全球资讯 · 微信公众号70

    AgiBot releases embodied Agent framework Genie Master; Genie G2 works and evolves in a convenience store

    AgiBot has released its embodied Agent framework Genie Master. Equipped with this framework, Genie G2 has taken a job in a real convenience store, where it does not rely on manual demonstrations or debugging; it autonomously identifies the causes of failures, adjusts its strategies, and accumulates experience, continuously improving on tasks such as packaged hot-food preparation, picking, and restocking, achieving self-evolution without a human in the loop.

  3. The Robot Report78

    WorldOracle releases action-conditioned world model benchmark; seven models show deviations between predictions and real execution

    Princeton University, UT Austin, the University of Toronto, NVIDIA, and AgiBot jointly launched WorldOracle, an action-conditioned world model benchmark that uses the same initial frame and action sequence to compare model predictions with real execution.

    Why it matters: The benchmark compares world model predictions frame by frame against real execution of the same actions and releases failure samples, making it easier to identify sources of model deviation over long sequences.

  4. 甲子光年 · 微信公众号45

    2026 Jiangsu AI+Science and Engineering Hackathon concludes: Agnes Harness supports 20 teams' AI4S prototypes

    The finals of the 2026 Jiangsu Province AI+Science and Engineering Innovation Practice Hackathon (University Group) have concluded. 536 teams registered and 20 advanced to the finals, with all projects built on Agnes Harness, the open-source agent development and execution infrastructure from Agnes.

  5. AI前线 · 微信公众号78

    Google DeepMind Releases World Model Genie 4, Supporting 24 FPS Real-Time Interaction and Native Audio

    Google DeepMind launched its next-generation world model Genie 4 on October 27 local time, calling it its most realistic and interactive world model to date, with a single generated interactive experience lasting up to 2 minutes, compared with 1 minute for the previous generation, Genie 3.

    Why it matters: Google is connecting language models' planning capabilities to real-time world generation, which lets readers understand a new approach to synthetic training grounds for embodied intelligence.

  6. 36氪 · 微信公众号84

    Google DeepMind releases world model Genie 4, supporting real-time interaction and native audio

    Google DeepMind released its next-generation general world model Genie 4 on the evening of the 28th. It can generate and modify interactive scenes in real time based on prompts, sustaining a single coherent world for about 2 minutes at roughly 24 frames per second.

    Why it matters: Genie 4 combines real-time world simulation, native audio, and multi-step reasoning, which lets readers judge that robot training environments are shifting from manual modeling to generative simulation.

  7. 硅星人Pro · 微信公众号78

    Google DeepMind releases world model Genie 4, integrated with Gemini for real-time generation and task reasoning

    Google DeepMind has released Genie 4, a new-generation world model that supports real-time generation of explorable virtual worlds and simultaneously generates physics effects, character movements, and native audio. Genie 4 is integrated with Gemini 3.8 Flash as a real-time guiding Agent, responsible for understanding goals, observing the environment, and planning next steps; in the demo, it simulated robotic arm operation in a virtual laboratory.

    Why it matters: Google released the Genie 4 world model, demonstrating its collaboration with Gemini to plan and simulate robot task execution in virtual environments, offering a look at a new way world models can be used for robot training.

  8. 机器之心 · 微信公众号58

    Chaoyan Intelligence releases 35B model Apex-35B-V1.0, leading Recursive Flint on 18 of 21 benchmarks

    Chaoyan Intelligence, founded by Chen Yongchao, an assistant professor at Tsinghua University's School of Artificial Intelligence, has released its first 35B model, Apex-35B-V1.0. It was trained and post-trained on Qwen3.6-35B-A3B, balancing general capabilities such as coding and agents with AI research, computational science, and robot simulation tasks.

  9. 机器之心 · 微信公众号73

    Google DeepMind releases world model Genie 4, supporting real-time control and native audio

    Google DeepMind has released Genie 4, its fourth-generation world model, which can generate dynamic environments in real time; users can control characters or the camera with the keyboard and use natural language to change the weather and add objects and characters. Genie 4 is Google's first world model to natively generate real-time audio, and it also improves physics simulation such as aerodynamics and collisions as well as the naturalness of character movements.

  10. 雷科技56

    Google releases Genie 4 world model: supports real-time interaction, can generate robot training scenarios

    Google has released a research preview of its world model Genie 4, which natively supports multimodality, can change the environment and introduce new objects and characters through instructions during generation, and lets existing characters react. The model is integrated with Gemini, with Gemini Omni anchoring the simulated visual behavior logic to the real world. It is currently open only to invited testers and will later open first to Ultra subscribers.

  11. 新智元 · 微信公众号80

    Google releases world model Genie 4, with three Gemini models jointly driving robot actions

    Google DeepMind has released Genie 4, a new-generation general-purpose world model that natively generates sound and improvised dialogue for the first time, sustaining a single coherent world for about 2 minutes at roughly 24 frames per second of real-machine footage. The model uses a hybrid world-model architecture: Gemini 3.8 Flash handles real-time control, while Gemini 4 Argon breaks tasks into 22-step action instructions, driving an Apollo robot to play checkers and, in simulation, controlling a UR5 robotic arm to perform pipetting operations.

    Why it matters: Google has released the Genie 4 world model, with three Gemini models jointly driving robot action decomposition; readers can learn how world models are used for robot training.

  13. 人形大讲堂 · 微信公众号62

    Three cross-embodiment papers examined: body, motion, and environment data still need manual translation

    The article examines three cross-embodiment research papers published between October 24 and 26, arguing that cross-embodiment is more like translation than a spontaneous model capability. MimicAll uses a structured body-condition table covering 16 series and 4,096 body variants.

  17. 中国机器人网 · 微信公众号66

    Xiaomi announces 21-DOF bionic hand, completing closed-loop validation based on 1,827 factory SOP data points

    Xiaomi has announced that its new-generation high-DOF bionic hand has completed closed-loop validation from scenario requirements to engineering design, with specifications of 21 DOF, 26-motor hybrid actuation, and 1,500-point full-palm tactile sensing. The development was based on motion-capture collection from 1,827 of 2,579 SOP steps across 16 work sections at the pilot center of Xiaomi Auto's intelligent manufacturing department. Simulation ablations show that the 17-to-21-DOF range improves average stable-grasp success rate by 15% to 38% over the 6-to-15 range.

  18. 量子位55

    Wika releases high-quality dataset for embodied intelligence-enabled construction machinery

    Wika Mobile Control released the "High-Quality Dataset for Embodied Intelligence-Enabled Construction Machinery" at the 10th Xuzhou Artificial Intelligence Conference 2026, and it is now available on data repositories including Hugging Face. The first phase of the dataset has 36 million frames, covers 12 typical operation scenarios, and synchronizes five dimensions of data—vision, LiDAR, pose, force feedback, and joystick input—using a unified timestamp. Wika says real-machine data collection for excavators costs about 5,000 yuan per hour, and training a usable model roughly requires data on the order of 100,000 hours.

  19. 极客公园63

    From Stanford Research to Red Sea Corals: How LingBot Models Are Entering the Foundation of Global Embodied AI

    Geek Park reviewed how the LingBot model series from Ant Group's embodied AI unit Robbyant has been adopted in academia, open-source communities, and industrial toolchains. Stanford's OpenWAM uses the LingBot-VA architecture as a reference implementation, and a KAIST team used LingBot-VLA as a teacher model to raise its student model's single-step inference success rate from 35.3% to 51.2%.

  20. 简智机器人 · X67

    GenRobot.AI's two open-source datasets enter Hugging Face robotics dataset trending top five

    GenRobot.AI announced that both of its open-source datasets, Gen-HumanWholeBody and Gen-HumanEgo, have entered the top five of the Hugging Face robotics dataset trending list. Gen-HumanWholeBody is a first-person whole-body manipulation dataset containing 600 hours of whole-body manipulation data, 16,976 real-world scenes, and 1,281 human skills, with whole-body MPJPE of about 3 cm; the authors say it targets whole-body motion learning, human-to-robot transfer, and embodied foundation model research.

    Why it matters: Two open-source datasets entering the top five of the Hugging Face robotics dataset trending list can serve as a reference for observing the level of attention to whole-body manipulation data.

  21. HuggingFace Daily Papers78

    RynnWorld-Latent: A Cross-Embodiment World Model Conditioned on Universal Latent Actions

    The research team proposes RynnWorld-Latent, a cross-embodiment video world model that uses the 3D-grounded latent actions of RynnLAM as a shared conditioning interface, enabling joint pretraining on human interaction videos, robot demonstrations, and simulation data.

    Why it matters: This work proposes using universal latent actions as a shared cross-embodiment conditioning interface, allowing human videos, robot demonstrations, and simulation data to jointly pretrain a world model.

  22. HuggingFace Daily Papers66

    PDRVLA: A Vision-Language-Action Model with Physical Dynamics Representations

    PDRVLA introduces a lightweight Physical Dynamics Module that extracts physical tokens from vision-language features to condition action generation, without requiring online future-video generation. The module learns by directly predicting future RGB, depth, and optical-flow features, with auxiliary supervision from a frozen pretrained video world model. Training uses only public data, expanded through human-to-robot synthesis, subtask annotation, and embodied chain-of-thought (ECoT) annotation.

  23. 芯东西73

    Google releases world model Genie 4, capable of simulating aerodynamics and collisions with native audio generation

    Google released its next-generation world model Genie 4 on October 28, capable of simulating physical processes such as aerodynamics and collisions, and natively generating music, ambient sound, and interactive dialogue. The model combines capabilities from the Gemini series, with Gemini 3.8 Flash breaking down multi-step tasks in a loop, and Google also demonstrated it combined with Gemini Robotics to simulate an Apptronik Apollo humanoid robot playing checkers.

  24. Guyuehome58

    How Much Does TEB Differ from DWA? 18 Gazebo Mecanum-Wheel Robot Simulations Compared

    The author ran 18 experiments comparing the TEB and DWA planners in a Gazebo mecanum-wheel robot simulation, with the title stating a 1.65x difference between them. The author also notes that when the planners moved sideways throughout, their lateral-movement rates differed by 100x, yet this was completely invisible on the trajectory plots.

  25. arXiv cs.RO62

    SCRIM: Quantifying State Criticality in Imitation Learning

    The authors propose SCRIM, a state criticality measure that uses an adversarial perturbation policy and a value critic to estimate the worst-case degradation in downstream value caused by small perturbations to expert actions, scoring each demonstration state. These scores can be used to reweight behavior-cloning losses or to select compact, high-value subsets of demonstrations, reducing training data while maintaining performance. The method is evaluated on a variety of simulated and real robot manipulation tasks, and the paper was accepted to CoRL 2026.

  26. arXiv cs.RO60

    Semantic World Models: Using Vision-Language Models to Predict Semantic Information for Robot Planning

    The paper proposes Semantic World Models, which model the world model as visual question answering about the semantic information of future frames rather than reconstructing future pixels, so that vision-language models can be obtained through supervised fine-tuning on image-action-text data. The authors validated it on simulated robot tasks, where it outperformed typical reconstruction-based action-conditioned world models.

  27. arXiv cs.RO72

    BiDex-100: A Bimanual Dexterous Manipulation Benchmark Covering 100+ Simulation Tasks

    The authors propose BiDex-100, a simulation benchmark for bimanual dexterous manipulation comprising 106 tasks organized along five capability dimensions: generalization, instruction following, limited demonstrations, long horizon, and memory. Each task is paired with an out-of-domain variant featuring controlled visual and spatial changes for evaluating robustness. The benchmark supports two embodiments and provides 8,080 success-validated teleoperated demonstrations totaling 6.25 million frames. The authors evaluate seven recent vision-language-action models under a unified protocol, establishing baselines for capability-specific performance, robustness, and data efficiency.

  29. arXiv cs.RO69

    IDTD: Information-Decoupled Trajectory Design for Sim-to-Real System Identification

    The authors propose an information-decoupled trajectory design framework, IDTD, which constructs an exploration objective from a Schur complement score derived from the Fisher information matrix, then normalizes by parameter, selects the optimal trajectory segment for each parameter, and aggregates the scores logarithmically, to mitigate unreliable identification caused by indistinguishable parameters.

  31. arXiv cs.RO66

    Learning-Based Informative Path Planning for Multi-Robot Systems: A Spatio-Temporal Gaussian Process Kalman Filter Framework

    The authors propose a grid-based spatio-temporal Gaussian process-Kalman filter framework for learning-based multi-robot informative path planning, representing anonymous target presence as a single latent variable field over a discrete workspace grid rather than maintaining one Gaussian process per target.

  32. arXiv cs.RO70

    From Grasping to Dexterous Manipulation: Large-Scale Grasp Pretraining for Dexterous Manipulation

    The research team proposes pretraining the low-level controller of a hierarchical imitation learning framework with large-scale dexterous grasp data, then fine-tuning it on downstream task demonstrations to support functional dexterity in articulated tool use. They built a 355k-trajectory grasp-pretraining dataset from large-scale dexterous grasp annotations and introduced DexCraft, a simulation benchmark with six articulated tool-use tasks. In simulation and real-world experiments, the method outperforms end-to-end diffusion policy baselines and hierarchical policies trained from scratch, improving real-world full-task success by 33.3 percentage points over DP3.

  33. arXiv cs.RO61

    LOPAL: Local Performance-Aware Active Learning from Imperfect Demonstrations

    LOPAL is a local performance-aware active learning method that uses a Gaussian Mixture Model (GMM) to jointly encode demonstration trajectories and their local quality assessments, thereby generating trajectories that outperform the imperfect demonstrations. The method also uses a shared autonomy (SA) mechanism to actively request user corrections in areas with insufficient data, while the robot autonomously executes its learned behavior during this process.

  34. arXiv cs.RO78

    FTP-1: A Generalist Foundation Tactile Policy Across Tactile Sensors

    The research team proposes FTP-1, a generalist foundation tactile policy pretrained across multiple tactile sensors and embodiments. It uses heterogeneous encoders to map image-, array-, and state-based tactile signals into unified morphology-aware latent tokens, which are jointly modeled by a shared tactile Transformer expert.

    Why it matters: This work unifies multiple tactile signals into a shared policy and reports changes in success rates for cross-sensor transfer, offering a reference for tactile manipulation research.

  35. arXiv cs.RO71

    RePO-VLA: Recovery-Driven Policy Optimization for Vision-Language-Action Models

    RePO-VLA proposes a recovery-driven policy optimization framework that assigns distinct roles to success, recovery, and failure trajectories for long-horizon, contact-rich manipulation tasks. The method first uses recovery-aware initialization to slice corrective segments into new episodes, then learns a progress-aware semantic value function, and uses value-conditioned refinement to make the policy prefer high-progress actions.

  37. arXiv cs.RO65

    MO-Playground: Massively Parallel Multi-Objective Reinforcement Learning for Robotics

    The authors propose MORLAX, a GPU-native multi-objective reinforcement learning algorithm, and MO-Playground, a pip-installable GPU-accelerated multi-objective environment library. They can approximate Pareto sets within minutes, achieving 26-271x speedups over traditional CPU methods and up to 19x speedups over prior GPU methods, while learning Pareto front hypervolumes that are superior. The team also implemented a custom BRUCE humanoid robot environment, learning Pareto-optimal locomotion policies in simulation across 6 practical objectives such as smoothness, energy efficiency, and arm swinging.

  40. arXiv cs.RO67

    DreamTacVLA: A VLA framework that uses a tactile world model to predict future contact

    DreamTacVLA proposes a framework that grounds VLA models in contact physics, learning contact dynamics by predicting future tactile signals. The model uses hierarchical perception, taking high-resolution tactile images as micro-vision inputs combined with wrist-camera local vision and third-person macro vision, and uses a Hierarchical Spatial Alignment (HSA) loss to align tactile tokens with their spatial counterparts in the wrist and third-person views.