Episodes

  • Hy-Embodied-0.5-VLA: A Massive Bimanual Teleoperation Dataset for Vision-Language-Action

    Hy-Embodied-0.5-VLA: A Massive Bimanual Teleoperation Dataset for Vision-Language-Action
    Jun 15 2026
    Released a massive bimanual robot manipulation dataset with 2,163 hours and 250K+ episodes across 70+ tasks, along with a compatible VLA model for multi-view egocentric teleop. The dataset and model are fully compatible with LeRobot v3.0.
    Show More Show Less
    21 mins
  • Q-Guided Flow: Test-Time Gradient Guidance of Flow Policies

    Q-Guided Flow: Test-Time Gradient Guidance of Flow Policies
    Jun 14 2026
    New framework for guided flow-matching policies that improves long-horizon robotic control and sample efficiency.
    Show More Show Less
    35 mins
  • Flow Reversal Steering: Guiding Diffusion-Based Robot Policies with High-Level Reasoning

    Flow Reversal Steering: Guiding Diffusion-Based Robot Policies with High-Level Reasoning
    Jun 14 2026
    Introduces flow reversal steering to guide diffusion-based vision-language-action models with high-level VLM reasoning and enables RL directly in the diffusion noise space.
    Show More Show Less
    38 mins
  • Test-Time Compute Scaling for Robot Policies (DIRECT)

    Test-Time Compute Scaling for Robot Policies (DIRECT)
    Jun 14 2026
    Larger models + more thinking + more context improve performance on some prompts but not others; a learned router enables better performance/latency trade-offs.
    Show More Show Less
    25 mins
  • LabVLA: Bringing Vision-Language-Action to the Chemistry Lab

    LabVLA: Bringing Vision-Language-Action to the Chemistry Lab
    Jun 14 2026
    RoboGenesis generates 10K+ lab scenes across 16 robot embodiments; LabVLA (Qwen3-VL + DiT flow-matching) achieves 71.1% success on LabUtopia and transfers to real Franka arms.
    Show More Show Less
    42 mins
  • Humanoid-GPT: A Foundation Model for Zero-Shot Humanoid Control

    Humanoid-GPT: A Foundation Model for Zero-Shot Humanoid Control
    Jun 13 2026
    GPT-style Transformer pretrained on 2 billion motion frames that achieves agile, generalist zero-shot control on a real Unitree G1 humanoid for tasks like soccer, dancing, and digging. Requires no fine-tuning or task-specific adaptation.
    Show More Show Less
    26 mins
  • CHORUS: Decentralized Multi-Robot Collaboration with a Single Shared VLA Model

    CHORUS: Decentralized Multi-Robot Collaboration with a Single Shared VLA Model
    Jun 13 2026
    Finetunes a single Vision-Language-Action (VLA) foundation model so that any robot in a team can control any other. Outperforms both per-robot specialists and a monolithic centralized policy while scaling to large teams.
    Show More Show Less
    37 mins
  • RISE: Self-Improving Robot Policy with Compositional World Model

    RISE: Self-Improving Robot Policy with Compositional World Model
    Jun 13 2026
    Trains a compositional world model on real robot data to enable closed-loop policy improvement via future prediction and progress evaluation, bypassing both risky real-world RL and traditional sim-to-real gaps.
    Show More Show Less
    40 mins