Skip to main content
Search

Beijing Academy of Artificial Intelligence

11 related articles

Subscribe to updates
  1. arXiv

    SLA moves spatial alignment into continual pre-training: 83.8% average success across eight real deployment conditions, 24.0 points over the baseline

    SLA extends spatial alignment for vision-language-action models from task post-training into continual pre-training, and keeps it during downstream post-training. Initialized from a pretrained policy, it jointly optimizes action prediction and a cosine alignment between intermediate visual embeddings and spatial representations from a frozen 3D perception model, using feature-dependent gates to aggregate multiple VLA layers and probabilistic state dropout to reduce reliance on demonstration trajectories. It reaches 72.3% on the RoboTwin 2.0 Clean2Random setting and 83.8% average success across eight real deployment conditions, 24.0 percentage points above the baseline.

  2. Splitting a humanoid's upper and lower body into two controllers cuts upper-body joint error to about 0.1 radians on H1 and H1-2

    Humanoid whole-body control usually drives the upper and lower body with one policy, so large arm motions can disturb balance and the two tasks interfere during training. JAEGER splits the body into two independent controllers that share observations and rewards, initializes the upper body with supervised learning, then trains the whole body with reinforcement learning while supporting both root-velocity and joint-angle command modes. In simulation comparisons on the H1 and H1-2 platforms, JAEGER's upper-body joint tracking error is lower than HumanPlus, ExBody and OmniH2O, with pose-mode joint errors about 30% to 60% of those methods.

  3. arXiv

    Two humanoids carry one payload through a narrow corridor: coordination becomes executable code, and a failed physical check rewrites roles

    When two humanoid robots carry one freely supported rigid payload through constrained corridors, team coordination is written as executable code rather than fixed by task decomposition, with a multimodal foundation model supplying motion primitives, semantic objectives, temporary roles and bounded parameters. A verifier judges each physical transition; rejected ones trigger structured revision from the latest accepted checkpoint, and verified experience is retained for reuse. Across 15 base-task runs with three foundation-model backbones, 8 tasks completed, with every backbone achieving end-to-end success.

  4. arXiv

    DiT features lag U-Net because noise rides the main path: long skip connections bypass it and lift mid-layer quality

    A controlled from-scratch comparison shows that adding long skip connections to a DiT improves its intermediate features across semantic correspondence, segmentation and depth, and the gain survives even when generation quality is matched by FID.

  5. arXiv

    Sub-agents can now ask to be reconfigured mid-task, lifting long-horizon success by 9.1%

    POrchestra lets sub-agents proactively send structured requests to the orchestrator during execution, so the orchestrator can revise their subgoal, context, tool set, or execution budget on the spot instead of waiting for the sub-agent to finish or fail. Earlier re-orchestration methods fixed each sub-agent's configuration until its lifecycle ended, delaying adaptation to dynamic environments. Across GAIA, GAIA2, and SWE-Bench Verified, task success improves by an average of 9.1% over the strongest baseline, and the companion OPOD turns that feedback into process supervision, raising small-orchestrator success by an average of 10.0% over training baselines.

  6. arXiv

    SlideDP speeds shared-host multi-GPU full-parameter fine-tuning by 1.46–2.64x and beats FSDP2's measured peak by 11.2% at a larger batch on four H100s

    SlideDP introduces a synchronous data-parallel runtime for shared-host multi-GPU systems that maintains one authoritative host state, decouples communication routes from state layout, and pipelines parameter delivery, gradient aggregation, and CPU updates across ranks and chunks, achieving geometric-mean throughput ratios of 1.46–2.64 over SlideFormer, MegaTrain, and ZeRO-Offload in matched-batch sweeps, approaching GPU-resident FSDP2 at a smaller batch and exceeding its measured peak throughput by 11.2% at a larger batch on four H100s, while supporting 256K-token sequences for Qwen3-14B and fine-tuning Qwen2.5-72B on four RTX 4090 GPUs.

  7. arXiv

    RoboCoach turns imagined failures into targeted demos: 150 extra subtask demonstrations lift Franka success from 13.3% to 75.0%

    The work presents RoboCoach, a world-model-guided active coaching framework whose Route-Imagine-Diagnose-Improve (RIDI) loop executes reusable skill experts in closed loop inside CoachWorld, a shared action-conditioned world model, uses a progress judge to record the first subtask that fails to complete, and thereby selects which subtask demonstrations to acquire and which expert adapters to update; across two simulation suites and two real-robot platforms, imagined and deployed success correlate with Spearman rho = 0.840 over 22 task-policy pairs, and with only 150 additional subtask demonstrations per platform success rises from 13.3% to 75.0% on Franka and from 40.0% to 83.8% on AgileX, while the coached experts reach 35.

  8. arXiv

    BiasRecon hits top PSNR in cross-anatomy, cross-center and cross-modality MRI reconstruction with fewer than 100 tunable parameters

    The work proposes BiasRecon, a bias-calibrated adaptation framework grounded in a minimal-intervention principle, which alternates frequency-guided prior calibration (learnable scalars α and β modulating low- and high-frequency components at each U-Net skip layer, adding only 2L parameters), score-based denoising, and adaptive regularization that uses Stein's Unbiased Risk Estimator (SURE) to tune the regularization parameter γ; trained on FastMRI knee (973 volumes) and evaluated without retraining on FastMRI-Brain (cross-anatomy), Stanford Knee (cross-center) and CMRxRecon (combined shift) under 8× Gaussian 1D undersampling, it improves PSNR over DDS by +1.19 dB, +2.06 dB and +1.60 dB respectively, and is best in 10 of 12 sampling/acceleration settings.

  9. arXiv

    MedQ-Engine lifts an 8B medical image quality model past GPT-4o by over 13% and cuts the gap to human experts to 4.34% using 10K annotations

    The work proposes MedQ-Engine, a closed-loop data engine that iterates through evaluate-explore-evolve phases: it clusters model failures on a development set into failure prototypes, uses the visual component of those prototypes as retrieval anchors over a roughly one-million-image pool spanning five modalities, and combines progressive human-in-the-loop annotation with entropy-guided routing and quality-assured fine-tuning, so that with only 10K annotations an 8B-parameter model surpasses GPT-4o by over 13% on medical image quality assessment and narrows the gap to human experts to 4.34%, with more than 4x sample efficiency over random sampling.

  10. arXiv

    UltraStar recasts echocardiography probe navigation from path regression to anchor-based global localization, beating baselines and scaling better with longer inputs on over 1.31 million samples

    The work proposes UltraStar, which reformulates echocardiography probe navigation from path regression to anchor-based global localization: a Star Graph treats historical keyframes as spatial anchors connected directly to the current view to explicitly model geometric constraints, and a semantic-aware sampling strategy actively selects representative landmarks from massive history logs to reduce redundancy for accurate anchoring; experiments on a dataset with over 1.31 million samples show it outperforms baselines and scales better with longer input lengths, indicating a more effective topology for history modeling under noisy exploration.