{
  "schema_version": "paper_public_manifest_v1",
  "paper_id": "tactile-glove-pose-artifacts_2026_07",
  "slug": "tactile-glove-pose-artifacts",
  "title": "Tactile glove readings are entangled with hand pose",
  "authors": [],
  "source": {
    "arxiv_id": "2607.22964",
    "pdf_url": "https://arxiv.org/pdf/2607.22964",
    "project_url": "",
    "github_url": "",
    "huggingface_url": "",
    "original_source": "https://arxiv.org/abs/2607.22964"
  },
  "site": {
    "post_url": "/posts/tactile-glove-pose-artifacts",
    "canonical_url": "https://haiguangboy.com/posts/tactile-glove-pose-artifacts",
    "cover_image": "https://static.haiguangboy.com/papers/tactile-glove-pose-artifacts/cover.webp"
  },
  "taxonomy": {
    "domain": "embodied_ai",
    "track": "world_model",
    "tasks": [
      "embodied_ai",
      "world_model",
      "vla",
      "robotics",
      "state_prediction",
      "Embodied intelligence",
      "Tactile perception",
      "Robot learning"
    ],
    "related_topics": [
      {
        "paper_id": "t_rex_tactile_reactive_dexterous_manipulation_2026_07",
        "title": "T-Rex: Why tactile sensing needs its own model",
        "url": "https://haiguangboy.com/posts/t_rex_tactile_reactive_dexterous_manipulation",
        "relation": "same_track",
        "summary": "T-Rex: Why tactile sensing needs its own model",
        "strength": "strong"
      },
      {
        "paper_id": "robointer15_a_holistic_intermediate_representation_suite_for_embodied_world_mode_2026_07",
        "title": "It's not \"whether to have a world model,\" it's \"whether what it outputs has structure\"",
        "url": "https://haiguangboy.com/posts/robointer15_a_holistic_intermediate_representation_suite_for_embodied_world_mode",
        "relation": "same_track",
        "summary": "It's not \"whether to have a world model,\" it's \"whether what it outputs has structure\"",
        "strength": "strong"
      },
      {
        "paper_id": "real_time_semg_based_telecontrol_of_an_assistive_robotic_arm_using_a_1d_convolut_2026_07",
        "title": "Neural signals: their value and their fragility come from the same place",
        "url": "https://haiguangboy.com/posts/real_time_semg_based_telecontrol_of_an_assistive_robotic_arm_using_a_1d_convolut",
        "relation": "same_track",
        "summary": "Neural signals: their value and their fragility come from the same place",
        "strength": "strong"
      },
      {
        "paper_id": "n_0_vtla_scaling_vision_tactile_language_action_model_with_latent_tactile_tokens_2026_07",
        "title": "Tactile sensing as a prediction target rather than an observation input",
        "url": "https://haiguangboy.com/posts/n_0_vtla_scaling_vision_tactile_language_action_model_with_latent_tactile_tokens",
        "relation": "same_track",
        "summary": "Tactile sensing as a prediction target rather than an observation input",
        "strength": "strong"
      }
    ]
  },
  "ruling": {
    "importance_score": 3.0,
    "one_sentence": "Tactile glove readings are entangled with hand pose"
  },
  "asset_base_url": "https://static.haiguangboy.com/papers/tactile-glove-pose-artifacts",
  "assets": [
    {
      "type": "cover_image",
      "object_key": "papers/tactile-glove-pose-artifacts/cover.webp",
      "content_type": "image/webp",
      "upload_status": "uploaded",
      "bucket": "paper-assets",
      "url": "https://static.haiguangboy.com/papers/tactile-glove-pose-artifacts/cover.webp",
      "role": "post_cover",
      "size_bytes": 88082
    },
    {
      "type": "public_brief",
      "object_key": "papers/tactile-glove-pose-artifacts/public_brief.md",
      "content_type": "text/markdown; charset=utf-8",
      "upload_status": "uploaded",
      "bucket": "paper-assets",
      "url": "https://static.haiguangboy.com/papers/tactile-glove-pose-artifacts/public_brief.md",
      "role": "public_brief",
      "size_bytes": 4710
    },
    {
      "type": "public_manifest",
      "object_key": "papers/tactile-glove-pose-artifacts/public_manifest.json",
      "content_type": "application/json; charset=utf-8",
      "upload_status": "uploaded",
      "bucket": "paper-assets",
      "url": "https://static.haiguangboy.com/papers/tactile-glove-pose-artifacts/public_manifest.json",
      "role": "public_manifest",
      "size_bytes": 4333
    }
  ],
  "published_at": "2026-07-29T22:27:41+08:00",
  "created_at": "2026-07-29T22:27:41+08:00",
  "updated_at": "2026-09-02T10:55:06+08:00",
  "analyst_take": {
    "type": "author_opinion",
    "text": "\"Don't directly splice auxiliary modalities into the backbone\"—this judgment has now accumulated its fifth independent source in the library, spanning five levels: VLA policies, tactile-specific architectures, world model latent variables, physiological signals, and now sensor calibration. The same pitfall being repeatedly hit at completely different levels of abstraction shows it's not a flaw of any one model class.\n\nThis paper's value lies in providing the mechanism: models tend to learn auxiliary modalities as direct predictors of the target quantity, rather than as correction terms to the main pathway. The fix is thus clear—not splicing them in, but giving them a structurally constrained position. Residual terms, prediction targets, and dedicated pathways are all different formulations of the same approach.\n\nThere is also a genuine disagreement worth noting: several tactile works in the library require specialized hardware, while this one advocates purely algorithmic retrofitting without modifying the glove. The hardware and algorithmic routes give opposite answers to \"at which layer should tactile data quality be addressed,\" and for now there is no verdict."
  }
}