{
  "schema_version": "1.1",
  "id": "s13:https://arxiv.org/abs/2610.32948v1",
  "slug": "2610-32948v1-0hbava6",
  "url": "https://feed7.dev/p/2610-32948v1-0hbava6",
  "title": "Beyond Discrete Tokens: Preserving Numerical Continuity in Medical Vision-Language Models",
  "why_included": "HAD-VLM combines text generation with diffusion-based continuous predictions for medical image measurements. On evaluated ultrasound and X-ray tasks, it reduced measurement error against the reported baselines.",
  "summary": "HAD-VLM pairs autoregressive text generation with diffusion-based prediction of measurement coordinates and lengths. Across four echocardiographic tasks, it reported **21.4–37.6% lower error** than a Vicuna-based autoregressive baseline; for X-ray vertebral measurement, error fell from **11.4 to 6.7 pixels**.",
  "practical_implication": "For builders handling images with precise numeric outputs, the paper is an example of separating language generation from continuous measurement prediction while sharing a model backbone. It also reports that medical question-answering performance was preserved on the evaluated dataset.",
  "agent_context": "HAD-VLM pairs autoregressive text generation with diffusion-based prediction of measurement coordinates and lengths. Across four echocardiographic tasks, it reported **21.4–37.6% lower error** than a Vicuna-based autoregressive baseline; for X-ray vertebral measurement, error fell from **11.4 to 6.7 pixels**.\n\nFor builders handling images with precise numeric outputs, the paper is an example of separating language generation from continuous measurement prediction while sharing a model backbone. It also reports that medical question-answering performance was preserved on the evaluated dataset.\n\nThe findings cover specific medical image tasks and datasets. The abstract does not establish performance beyond those evaluations or provide deployment details.",
  "source": {
    "name": "arXiv",
    "url": "https://arxiv.org/abs/2610.32948v1",
    "published_at": "2026-10-23T17:40:58.000Z"
  },
  "source_class": "blog_post",
  "content_type": "Paper",
  "layer": "model",
  "domains": [
    "image",
    "research"
  ],
  "topics": [],
  "verification": {
    "status": "needs_review",
    "label": "Needs Review",
    "method": "unverified",
    "verified_at": null
  },
  "uncertainty": [
    "The findings cover specific medical image tasks and datasets. The abstract does not establish performance beyond those evaluations or provide deployment details."
  ],
  "connected_context": {
    "meaning": "This narrows multimodal medical measurement toward a split output design: language remains autoregressive while coordinates and lengths use diffusion, with gains reported on the evaluated tasks. It complements broader medical image understanding work with a specific approach to continuous measurements; the evidence does not establish performance beyond those datasets.",
    "corpus_size": 910,
    "generated_at": "2026-10-26T11:01:08.586Z",
    "connections": [
      {
        "title": "ClinFusion: A Vision-Centric Multimodal LLM System for Holistic Medical Understanding",
        "source_name": "arXiv",
        "source_url": "https://arxiv.org/abs/2607.24743v1",
        "feed7_url": "https://feed7.dev/p/2607-24743v1-15vx95s",
        "reason": "Both address medical image models, but ClinFusion focuses on holistic 2D and 3D understanding while HAD-VLM separates precise continuous measurement from text generation."
      }
    ]
  },
  "lifecycle": "Current",
  "published_at": "2026-10-23T17:40:58.000Z",
  "modified_at": "2026-10-23T17:40:58.000Z",
  "supersedes": [],
  "expires_at": null,
  "formats": {
    "html": "https://feed7.dev/p/2610-32948v1-0hbava6",
    "json": "https://feed7.dev/p/2610-32948v1-0hbava6.json",
    "markdown": "https://feed7.dev/p/2610-32948v1-0hbava6.md"
  }
}