庞玉栋博客
//

计算机视觉

21 篇 · 资讯流
  2. arXiv cs.CL（自然语言）

    VisionWeave: Weaving Elastic Visual Representations as a Native Capability of MLLMs

    Multimodal large language models have become the dominant paradigm for visual understanding, but incur substantial costs

    原文 ↗
  3. arXiv cs.RO（机器人）

    EmbodiedSmith: Scaling Embodied Data through Recursive Self-Improvement Flywheel in Simulation

    Scaling robotic foundation models requires diverse training data and reliable evaluation environments. Simulation offers

    原文 ↗
  5. arXiv cs.RO（机器人）

    GroundingPI: A Grounding Foundation Model towards Physical Intelligence with Visual Primitives

    Precise grounding matters. It specifies which object is the target and where that object is, even in clutter and for tin

    原文 ↗
  6. arXiv cs.RO（机器人）

    GroundAnything: Reconciling Parallel Decoding with Precise Visual Grounding at Flash Speed

    Autoregressive (AR) grounding models serialize spatial predictions, introducing sequential latency and imposing a causal

    原文 ↗
  7. arXiv cs.AI（人工智能）

    $S^3$: Spectral Null-Space Swap Makes Reasoning Models Efficient

    LLMs trained with Chain-of-thought excel in reasoning capability, but often come with excessive token cost. We find that

    原文 ↗
  9. arXiv cs.RO（机器人）

    JRDB-AVR: An Active Visual Reasoning Benchmark for Embodied Agents in Real-World Environments

    In complex embodied visual reasoning scenarios, an agent often has only a limited field of view, and the evidence needed

    原文 ↗
  11. arXiv cs.CV（计算机视觉）

    M3GD: Multi-Modal Multi-View Geometric Diffusion for Camera--LiDAR Novel View Synthesis

    Robotic novel view synthesis (NVS) must recover both visual appearance and metric 3D structure, yet most generative NVS

    原文 ↗
  13. arXiv cs.CV（计算机视觉）

    Retrieve-to-Localize: Bridging Large Language Models and LiDAR Geometry for Spatial Grounding

    LiDAR provides precise geometric information for spatial perception tasks such as object detection in autonomous driving

    原文 ↗
  15. arXiv cs.CV（计算机视觉）

    ForeDrive: Foresight-Guided End-to-End Autonomous Driving with a Planning-Relevant Latent World Model

    Existing latent world models are typically optimized for future predictability, yet the resulting representations are no

    原文 ↗
  16. arXiv cs.CV（计算机视觉）

    Metric-Bench: Exploring In-context Spatial Metric Reasoning in VLMs for Indoor Scenes

    Metric reasoning is a critical and challenging task for Vision Language Models (VLMs), playing a pivotal role in embodie

    原文 ↗