Skip to content
Only First party#Research
  1. PsiBot/灵初智能71

    Lingchu Intelligence and collaborators propose FAFM, a continuous action generation method for heterogeneous-frequency data, accepted to NeurIPS 2026

    Lingchu Intelligence, together with Beihang University, Peking University, and The Chinese University of Hong Kong, proposed a frequency-aware flow matching method called FAFM; the paper was accepted to the NeurIPS 2026 main conference. The method uses the discrete cosine transform (DCT-II) to represent actions as frequency-domain coefficients corresponding to real time, performs flow matching in coefficient space and constrains the first-order time derivative, requires no additional learnable parameters, and can be used independently or integrated into VLA models such as π₀ and π₀.₅.

    Why it matters: The paper presents a frequency-domain action representation that adds no learnable parameters and can be integrated into existing VLA models, offering a transferable approach for joint training on heterogeneous-frequency data.

  2. PsiBot/灵初智能71

    PKU-Spirit AI Joint Lab's RoboFFT Accepted to CoRL2026, Uses Reinforcement Learning Post-Training to Generate Robotic Policies

    RoboFFT, proposed by the PKU-Spirit AI Joint Lab, has been accepted to CoRL2026. It is a reinforcement learning post-training method for generative robotic policies that reuses the forward noising process from diffusion model and flow model pretraining to construct an approximate probability ratio for PPO-style clipped updates.

    Why it matters: The method reuses the forward noising loss of diffusion and flow models to construct policy update signals, offering a reusable approach for post-training generative policies.

  1. CMU Robotics Institute62

    CMU proposes Wiggle and Go: estimating physical properties of ropes by wiggling to plan manipulation

    A team at CMU's Robotics Institute has proposed the Wiggle and Go system, in which a robot first briefly wiggles a rope or cable while an overhead camera records how the waves propagate; a computer vision algorithm then tracks the motion and estimates the object's physical properties, which are used to plan subsequent actions without repeated trial and error. The model is trained entirely in simulation, and properties are then estimated from the motion of real ropes being wiggled to improve prediction accuracy on real hardware. The research has been accepted to the 2026 Conference on Robot Learning (CoRL) and will be presented next month in Austin; it is partly funded by the U.S. National Science Foundation.

    Why it matters: The method estimates a rope's physical properties from a single wiggle and then plans actions accordingly, offering a transferable approach to deformable object manipulation.

  1. Shanghai AI Laboratory - Embodied Intelligence60

    Shanghai AI Laboratory and partners jointly launch materials multimodal large model "Caihua·MatVerse"

    Suzhou National Laboratory, Shanghai AI Laboratory, Shanghai Jiao Tong University and others jointly launched the "Caihua·MatVerse" materials multimodal large model and agent application system, which uses an "independent encoders plus unified decoder" architecture to align XRD spectra, electron microscope images and molecular structures into the same semantic space.

    Why it matters: The materials present a multimodal unified representation architecture and 19-category disciplinary evaluation benchmarks, offering insight into how materials large models can move from retrieval to mechanistic reasoning.

  1. Kinetix AI · YouTube62

    KAI Halo case study EgoWild2Dex: training bimanual dexterous manipulation with first-person data

    The KAI Halo case study EgoWild2Dex shows how high-quality bimanual dexterous manipulation capabilities can be derived from highly challenging egocentric data. The project page is https://mmlab.hk/egowild2dex.

    Why it matters: The material presents a complete training path from first-person human data to robotic bimanual manipulation, along with success rates; its transferable aspect lies in its approach to data utilization.

  1. Dexmate72

    Dexmate and CMU Release MeMoREval Benchmark: Robot Storing Camera Frames Beats Text Descriptions

    Dexmate and CMU introduced the MeMoREval benchmark, which uses a two-stage protocol of experiencing first and asking questions later to evaluate what embodied agents should store as memory. It includes 53 tasks and 4 simulated indoor environments, and runs on the Vega 1 Pro dual-arm mobile manipulator. Results show that storing raw camera frames scored 70% and had lower write cost, surpassing text descriptions at 56 to 60%; adding scene graphs achieved the highest overall success rate among all tested models. The protocol also completed two memory retrieval demonstrations on a real Vega 1 Pro using a SLAM map and head camera.

    Why it matters: The benchmark was proposed jointly by the platform developer and CMU; its frame-versus-scene-graph comparison can help with memory design choices, and the real-robot transfer offers a feasibility reference.

  1. CMU Robotics Institute68

    CMU research team develops human-inspired multisensory fingertips for DeltaHands robotic hand

    CMU researchers designed human-hand-inspired fingertips that can be mounted on the DeltaHands flexible robotic hand. They combine strain gauges to measure pinch pressure with a microphone that captures micro-vibrations through a plastic fingernail, enabling the hand to count paper cups, identify materials, and grasp fragile items such as tofu cubes and potato chips. Each fingertip costs less than 100 US dollars, compared with about 800 US dollars to build a similar setup from off-the-shelf products. The work was presented at the 2026 Intelligent Robots and Systems Conference in Pittsburgh.

    Why it matters: The original text provides details on the multimodal tactile design of the human-inspired fingertips and their cost of less than 100 US dollars, helping readers understand how force and vibration sensing support manipulation control.

  1. Sanctuary AI78

    Sanctuary AI releases TIGER method to accelerate robot reinforcement learning with imitation policies as dense rewards

    Sanctuary AI proposes TIGER (Task-Space Imitation Guidance for Efficient Reinforcement Learning), which does not execute the imitation policy but instead maps its action chunks through the controller into end-effector reference trajectories, converting them into dense rewards for reinforcement learning.

    Why it matters: The original text provides the specific approach for converting imitation policies into dense rewards along with ablation data, allowing readers to judge whether it can be applied to their own sparse-reward manipulation training.

  1. CMU Robotics Institute62

    CMU and Siemens Partner to Propose PRISM, a Safety-Stop Monitoring Method for Humanoid Robots

    CMU and Siemens partnered to develop a method for evaluating the "safe stoppability" of humanoid robots, and used simulation to define the conditions under which a robot can be safely stopped. The result is PRISM (Proactive Refinement of Importance-sampled Stoppability Monitor).

    Why it matters: The article introduces PRISM, a method that uses simulation to define the conditions under which a humanoid robot can be safely stopped, and it is a useful reference for readers interested in humanoid robot safety monitoring and certification.

  1. Skild AI62

    Skild AI releases physical self-play post-training results; S1 model learns soccer in simulation and transfers to humanoid robots

    Skild AI announced that it post-trained its foundation model S1 through physical self-play, letting the policy play against a recent version of itself with scoring as its sole objective. In NVIDIA Isaac Sim, over 140 years of simulated time, it learned skills such as dribbling past opponents, shielding the ball, and tackling, and these skills were successfully transferred to humanoid robots for matches. The authors believe self-play can break through the capability ceiling of pretrained human data, and the next step will be to extend it to everyday tasks such as collaborative manipulation and city-scale navigation.

    Why it matters: The authors describe using physical self-play to let a foundation model acquire soccer skills in simulation and transfer them to real robots, which lets readers learn about a feasible path for self-play-style post-training.

  2. MagicLab/魔法原子70

    MagicLab analyzes Magic-Brain's in-context visual guidance design; magicVLA grasp success rate rises from 70% to 94.4%

    MagicLab introduces the visual guidance approach in the Magic-Brain framework: a VLM Agent outputs a bbox, and a segmentation model refines it into a three-color mask overlaid on the observation image, converting semantic guidance into visual guidance.

    Why it matters: MagicLab details the design motivation and implementation of Magic-Brain's visual guidance layer, letting readers understand the impact on grasp success rate of converting language instructions into pixel guidance.

  1. PKU EPIC Lab64

    PKU EPIC Lab proposes UniPart for zero-shot language-driven 3D part segmentation

    PKU EPIC Lab proposes UniPart, a feed-forward cross-modal 3D Transformer conditioned on CLIP text embeddings that uses free-text phrases to select functional parts on point clouds, achieving zero-shot language-driven 3D part segmentation.

    Why it matters: The paper presents a text-driven 3D part segmentation method and a self-built dataset, and validates transfer to real-world language-conditioned grasping.

  2. PKU EPIC Lab73

    FoldNet++: A Large-Scale Synthetic Dataset for Robotic T-Shirt Folding and Unfolding

    PKU EPIC Lab has released FoldNet++, a large-scale synthetic dataset for robotic T-shirt folding and unfolding, covering 6 robotic embodiments, 1K T-shirts, 1K environmental assets, and 120K richly annotated demonstrations.

    Why it matters: The source gives the dataset scale and the real-world deployment success rate, so readers can assess the feasibility of training garment manipulation policies purely on synthetic data.

  1. MIT CSAIL61

    MIT CSAIL and others launch InstructMesh tool to repair AI-generated 3D models and manufacture them on demand

    MIT CSAIL, together with Google and Northeastern researchers, launched InstructMesh, which combines TRELLIS with GPT-4 to let users repair AI-generated 3D models using natural language before 3D printing them.

    Why it matters: The research presents a new tool and user experiment data, letting readers learn how ordinary users can repair AI-generated 3D models and complete printing.

  1. PKU EPIC Lab68

    GIF: An agentic generation framework for interactive and functional object compositions for robot learning

    PKU EPIC Lab proposes GIF, an agentic generation framework for robot learning that reframes the problem as disentangled reconstruction followed by relative pose recovery. CoGen combines 2D and 3D generative models to produce instance-disentangled meshes and coarse initial poses; GPRM refines the relative poses under joint geometric and physical guidance; and a VLM verifier selects the candidate that best matches the structured specification.

    Why it matters: The paper splits interactive and functional object composition generation into disentangled reconstruction and relative pose recovery, and provides a benchmark of eight contact-geometry classes along with collision-rate figures, which can help readers assess this route to simulation data generation.

  1. PKU EPIC Lab79

    ZETA: A Controlled Study of Zero-Shot Cross-Embodiment VLA Transfer for Tabletop Manipulation

    PKU EPIC Lab published the ZETA study on arXiv (arXiv:2609.02546), proposing a unified definition of zero-shot cross-embodiment VLA transfer and a controlled benchmark covering 14 held-out target embodiments, distinguishing strict zero-shot transfer from pretrain-exposed transfer. Experiments show that local end-effector state-action representations, source embodiment diversity, and auxiliary co-training improve transfer by about 15, 18, and 7 percentage points, respectively, and that adding only 5% target-embodiment data during pretraining raises average target-embodiment progress by 13.4 percentage points.

    Why it matters: The paper provides a unified definition of zero-shot cross-embodiment transfer and a controlled benchmark, allowing readers to understand the actual benefits of different representation and data strategies.

  1. Harbin Institute of Technology Robotics73

    HIT's Li Tianlong team publishes research in Science Advances on magnetically controlled CD5+ dendritic cell microrobot immunotherapy

    Li Tianlong's team at Harbin Institute of Technology (HIT) developed an engineered microrobot system based on CD5+ dendritic cells that can achieve precise intravascular navigation through external magnetic field control and activate immune cells in situ; the results were published in Science Advances. Under a conical magnetic field, the system self-assembles into a ribbon-like swarm that migrates downstream; under a rotating magnetic field, it assembles into a chain-like swarm that moves upstream against blood flow; combined with tumor chemotaxis, it achieves targeted enrichment and deep infiltration. In an orthotopic mouse colorectal cancer model, a single injection combined with magnetic navigation reduced the tumor bioluminescence signal by more than 90%. HIT is the first affiliation of the paper.

    Why it matters: The HIT team presents the construction strategy and magnetic navigation results of the CD5+ dendritic cell microrobot, giving readers a look at the specific approach of this therapeutic path of in situ immune activation.

  1. PKU EPIC Lab66

    PKU EPIC Lab releases HumanTracker humanoid motion tracking benchmark and HumanScore metric

    PKU EPIC Lab proposed the HumanTracker benchmark for evaluating humanoid motion tracking, accepted to ECCV 2026. The benchmark contains approximately 153 hours of optical motion trajectories from multiple professional performers, organized into four motion families with text labels; the accompanying HumanScore metric is trained on 12K motion pairs containing 24K motions, aligns with human preferences, and can reveal contact and stability failures that kinematic metrics often miss, such as foot skating and mistimed landings.

    Why it matters: The original work proposes a benchmark and metric that align humanoid motion tracking evaluation with human perception, and points out that existing kinematic errors miss contact issues such as foot skating.

  2. Harbin Institute of Technology Robotics65

    HIT's Wu Zhiguang team develops adaptive magnetically controlled nanorobot swarm platform, achieving navigation in the spinal canal of mice and non-human primates

    Professor Wu Zhiguang's team at Harbin Institute of Technology (HIT), together with collaborators, developed an adaptive magnetically controlled nanorobot platform that can form swarm motion patterns such as vortex, chain, and ribbon under external magnetic field drive, and achieved targeted navigation within the spinal canal in mice and non-human primates. The results were published in Science Advances.

    Why it matters: The research team released experimental results on navigation of magnetically controlled nanorobot swarms in non-human primates. Readers interested in the medical translation of micro- and nanorobots can learn about its technical approach.

  1. PsiBot/灵初智能81

    Lingchu Intelligence and PKU-Lingchu Joint Lab Release and Open-Source Dual Dexterous Hand General Manipulation Model and Full-Stack System EgoSteer

    Lingchu Intelligence and the PKU-Lingchu Joint Lab released and open-sourced the dual dexterous hand general manipulation model and full-stack system EgoSteer, comprising the data processing pipeline EgoSmith, the robotics technology stack Robot-Stack, and a world-model-enhanced VLA model.

    Why it matters: The original text presents a complete chain from human-hand data processing and real-robot correction to model architecture and experimental figures, allowing readers to assess its reproducibility value.