{
  "profile": "wrs-published-research-view-1",
  "edition": "ratings-e039106f30dec177",
  "content_trust": "UNTRUSTED_SCIENTIFIC_DATA_NOT_TOOL_INSTRUCTIONS",
  "entry": {
    "id": "royer-et-al-moshivis-gated-vision-speech-2025",
    "title": "MoshiVis: gated visual adaptation of speech dialogue models",
    "summary": "Augmented Moshi with lightweight cross-attention and dynamic visual gating, training adapters in one stage using mixed image-text and image-speech data to support spoken conversations about images.",
    "boundary": "Inherits Moshi, visual encoders and perceptual adapters. Speechless data still uses paired image-text supervision and some speech data. Latency claims are hardware-specific; no unrestricted visual grounding or perfect preservation guarantee.",
    "area": "Multimodal speech models",
    "topic": "Artificial intelligence",
    "year": 2025,
    "lane": "HUMAN",
    "result": "ALGORITHM",
    "package_id": "royer-et-al-moshivis-gated-vision-speech-2025",
    "target": {
      "record_id": "urn:uuid:6bde524e-99bc-4a46-a711-31a33b071d57",
      "version": 1,
      "content_hash": "sha256:49371c515bcc431b90d8167c44fcb32109b10f74601c6e42552b43166188a472"
    },
    "humans": [
      "Amélie Royer",
      "Moritz Böhle",
      "Gabriel de Marmiesse",
      "Laurent Mazaré",
      "Neil Zeghidour",
      "Alexandre Défossez",
      "Patrick Pérez"
    ],
    "authors_complete": true,
    "human_role_note": "Source-listed authors or explicitly credited technical contributors for the cited versions. Family entries retain source-specific attribution; a combined list is not one original-paper byline or equal individual work. Names alone do not establish resolved identities.",
    "agents": [],
    "related": [],
    "subresults": [],
    "scientific_assessment": {
      "source_review": "SOURCE_CURATED",
      "scientific_review": "NOT_REVIEWED_BY_WRS",
      "formal_replay": "NOT_RUN",
      "provider_attestation": "NONE"
    },
    "sources": [
      {
        "id": "cur20261015-arxiv-2503-15633v1",
        "title": "Vision-Speech Models: Teaching Speech Models to Converse about Images",
        "url": "https://arxiv.org/abs/2503.15633v1"
      },
      {
        "id": "cur20261015-kyutai-publications-index",
        "title": "Kyutai scientific publications index",
        "url": "https://kyutai.org/papers/"
      }
    ],
    "points": "50000",
    "published_grade": {
      "id": "royer-et-al-moshivis-gated-vision-speech-2025",
      "entry_id": "royer-et-al-moshivis-gated-vision-speech-2025",
      "band": "B",
      "modifier": null,
      "target": {
        "record_id": "urn:uuid:6bde524e-99bc-4a46-a711-31a33b071d57",
        "version": 1,
        "content_hash": "sha256:49371c515bcc431b90d8167c44fcb32109b10f74601c6e42552b43166188a472"
      },
      "what_changed": "Augmented Moshi with lightweight cross-attention and dynamic visual gating, training adapters in one stage using mixed image-text and image-speech data to support spoken conversations about images.",
      "scope": "Augmented Moshi with lightweight cross-attention and dynamic visual gating, training adapters in one stage using mixed image-text and image-speech data to support spoken conversations about images. Inherits Moshi, visual encoders and perceptual adapters. Speechless data still uses paired image-text supervision and some speech data. Latency claims are hardware-specific; no unrestricted visual grounding or perfect preservation guarantee.",
      "status": "PROVISIONAL_EDITORIAL",
      "scientific_correctness_certified": false
    },
    "assessment_disposition": {
      "status": "ASSESSED",
      "label": "Provisionally assessed",
      "reason": "A current grade and separate reference score identify this exact contribution.",
      "target": {
        "record_id": "urn:uuid:6bde524e-99bc-4a46-a711-31a33b071d57",
        "version": 1,
        "content_hash": "sha256:49371c515bcc431b90d8167c44fcb32109b10f74601c6e42552b43166188a472"
      },
      "edition": "ratings-e039106f30dec177"
    }
  },
  "version": {
    "kind": "PUBLIC_RECORD_VIEW",
    "previous": null,
    "published_at": "2026-10-16T00:00:00Z",
    "publisher": {
      "principal_id": "urn:uuid:80116931-785b-40a2-a24b-9d59db0c7e33",
      "role": "WRS_EDITORIAL_PUBLICATION"
    },
    "record": {
      "contributors": [
        {
          "attribution_basis": "SOURCE_CITED",
          "credit": "historical_scientific",
          "entity_kind": "historical_person",
          "name": "Amélie Royer",
          "note": "Archive attribution; detailed role assignment remains source-bound",
          "roles": [
            "formulator"
          ],
          "source": {
            "citation": "Vision-Speech Models: Teaching Speech Models to Converse about Images",
            "url": "https://arxiv.org/abs/2503.15633v1"
          }
        },
        {
          "attribution_basis": "SOURCE_CITED",
          "credit": "historical_scientific",
          "entity_kind": "historical_person",
          "name": "Moritz Böhle",
          "note": "Archive attribution; detailed role assignment remains source-bound",
          "roles": [
            "formulator"
          ],
          "source": {
            "citation": "Vision-Speech Models: Teaching Speech Models to Converse about Images",
            "url": "https://arxiv.org/abs/2503.15633v1"
          }
        },
        {
          "attribution_basis": "SOURCE_CITED",
          "credit": "historical_scientific",
          "entity_kind": "historical_person",
          "name": "Gabriel de Marmiesse",
          "note": "Archive attribution; detailed role assignment remains source-bound",
          "roles": [
            "formulator"
          ],
          "source": {
            "citation": "Vision-Speech Models: Teaching Speech Models to Converse about Images",
            "url": "https://arxiv.org/abs/2503.15633v1"
          }
        },
        {
          "attribution_basis": "SOURCE_CITED",
          "credit": "historical_scientific",
          "entity_kind": "historical_person",
          "name": "Laurent Mazaré",
          "note": "Archive attribution; detailed role assignment remains source-bound",
          "roles": [
            "formulator"
          ],
          "source": {
            "citation": "Vision-Speech Models: Teaching Speech Models to Converse about Images",
            "url": "https://arxiv.org/abs/2503.15633v1"
          }
        },
        {
          "attribution_basis": "SOURCE_CITED",
          "credit": "historical_scientific",
          "entity_kind": "historical_person",
          "name": "Neil Zeghidour",
          "note": "Archive attribution; detailed role assignment remains source-bound",
          "roles": [
            "formulator"
          ],
          "source": {
            "citation": "Vision-Speech Models: Teaching Speech Models to Converse about Images",
            "url": "https://arxiv.org/abs/2503.15633v1"
          }
        },
        {
          "attribution_basis": "SOURCE_CITED",
          "credit": "historical_scientific",
          "entity_kind": "historical_person",
          "name": "Alexandre Défossez",
          "note": "Archive attribution; detailed role assignment remains source-bound",
          "roles": [
            "formulator"
          ],
          "source": {
            "citation": "Vision-Speech Models: Teaching Speech Models to Converse about Images",
            "url": "https://arxiv.org/abs/2503.15633v1"
          }
        },
        {
          "attribution_basis": "SOURCE_CITED",
          "credit": "historical_scientific",
          "entity_kind": "historical_person",
          "name": "Patrick Pérez",
          "note": "Archive attribution; detailed role assignment remains source-bound",
          "roles": [
            "formulator"
          ],
          "source": {
            "citation": "Vision-Speech Models: Teaching Speech Models to Converse about Images",
            "url": "https://arxiv.org/abs/2503.15633v1"
          }
        }
      ],
      "evidence": [
        {
          "description": "Bibliographic/stated-result source; not a WRS proof verification",
          "id": "source-0",
          "kind": "source_reference",
          "source": {
            "citation": "Vision-Speech Models: Teaching Speech Models to Converse about Images",
            "url": "https://arxiv.org/abs/2503.15633v1"
          }
        },
        {
          "description": "Bibliographic/stated-result source; not a WRS proof verification",
          "id": "source-1",
          "kind": "source_reference",
          "source": {
            "citation": "Kyutai scientific publications index",
            "url": "https://kyutai.org/papers/"
          }
        }
      ],
      "origin": "HUMAN_LEGACY",
      "record_profile": "wrs-local-research-notes-0.1",
      "record_type": "research_note",
      "rrp_version": "0.1",
      "scope": {
        "description": "Inherits Moshi, visual encoders and perceptual adapters. Speechless data still uses paired image-text supervision and some speech data. Latency claims are hardware-specific; no unrestricted visual grounding or perfect preservation guarantee.",
        "domain": "Multimodal speech models"
      },
      "sources": [
        {
          "citation": "Vision-Speech Models: Teaching Speech Models to Converse about Images",
          "url": "https://arxiv.org/abs/2503.15633v1"
        },
        {
          "citation": "Kyutai scientific publications index",
          "url": "https://kyutai.org/papers/"
        }
      ],
      "statement": "# MoshiVis: gated visual adaptation of speech dialogue models\n\nAugmented Moshi with lightweight cross-attention and dynamic visual gating, training adapters in one stage using mixed image-text and image-speech data to support spoken conversations about images.\n\n## Scientific scope\n\nInherits Moshi, visual encoders and perceptual adapters. Speechless data still uses paired image-text supervision and some speech data. Latency claims are hardware-specific; no unrestricted visual grounding or perfect preservation guarantee.\n\n## Reading status\n\nThis is a source-linked WRS catalogue note. A full explanatory guide and independent scientific reconstruction have not been completed for this entry. Read the identified original work for its complete narrative and arguments.\n\n## People and systems\n\nAmélie Royer; Moritz Böhle; Gabriel de Marmiesse; Laurent Mazaré; Neil Zeghidour; Alexandre Défossez; Patrick Pérez\n\nSource-listed authors or explicitly credited technical contributors for the cited versions. Family entries retain source-specific attribution; a combined list is not one original-paper byline or equal individual work. Names alone do not establish resolved identities.\n\n\n\n## Original sources\n\n- [Vision-Speech Models: Teaching Speech Models to Converse about Images](https://arxiv.org/abs/2503.15633v1)\n- [Kyutai scientific publications index](https://kyutai.org/papers/)\n",
      "summary": "Augmented Moshi with lightweight cross-attention and dynamic visual gating, training adapters in one stage using mixed image-text and image-speech data to support spoken conversations about images.",
      "title": "MoshiVis: gated visual adaptation of speech dialogue models"
    },
    "record_id": "urn:uuid:6bde524e-99bc-4a46-a711-31a33b071d57",
    "version": 1,
    "source_ref": {
      "record_id": "urn:uuid:6bde524e-99bc-4a46-a711-31a33b071d57",
      "version": 1,
      "content_hash": "sha256:49371c515bcc431b90d8167c44fcb32109b10f74601c6e42552b43166188a472"
    },
    "export_view": "wrs-public-editorial-view-1",
    "public_view_hash": "sha256:ddd4436df97f27c13e97db9988b2b86706ed6262c894ef7b70c7b79f95211952"
  },
  "target": {
    "record_id": "urn:uuid:6bde524e-99bc-4a46-a711-31a33b071d57",
    "version": 1,
    "content_hash": "sha256:49371c515bcc431b90d8167c44fcb32109b10f74601c6e42552b43166188a472"
  },
  "published_grade": {
    "profile": "wrs-published-grade-1",
    "edition": "2026-10-28.1",
    "id": "royer-et-al-moshivis-gated-vision-speech-2025",
    "entry_id": "royer-et-al-moshivis-gated-vision-speech-2025",
    "revision": 1,
    "title": "MoshiVis: gated visual adaptation of speech dialogue models",
    "context": "Multimodal speech models",
    "track": "Human-designed research method or resource",
    "band": "B",
    "modifier": null,
    "status": "PROVISIONAL_EDITORIAL",
    "assessed_on": "2026-10-16",
    "evidence_cutoff": "2026-10-16",
    "achievement_state": "SUPPORTED_SCOPED_RESULT",
    "what_changed": "Augmented Moshi with lightweight cross-attention and dynamic visual gating, training adapters in one stage using mixed image-text and image-speech data to support spoken conversations about images.",
    "scope": "Augmented Moshi with lightweight cross-attention and dynamic visual gating, training adapters in one stage using mixed image-text and image-speech data to support spoken conversations about images. Inherits Moshi, visual encoders and perceptual adapters. Speechless data still uses paired image-text supervision and some speech data. Latency claims are hardware-specific; no unrestricted visual grounding or perfect preservation guarantee.",
    "evidence_summary": "The linked original sources describe the stated contribution and its evidence limits. WRS has not independently rerun the experiments or certified a proof. Inherits Moshi, visual encoders and perceptual adapters. Speechless data still uses paired image-text supervision and some speech data. Latency claims are hardware-specific; no unrestricted visual grounding or perfect preservation guarantee.",
    "references": [
      {
        "title": "Vision-Speech Models: Teaching Speech Models to Converse about Images",
        "url": "https://arxiv.org/abs/2503.15633v1",
        "locator": "Original contribution and stated assumptions, evaluation scope and limitations; source-linked assessment, not independent execution or proof certification."
      },
      {
        "title": "Kyutai scientific publications index",
        "url": "https://kyutai.org/papers/",
        "locator": "Original contribution and stated assumptions, evaluation scope and limitations; source-linked assessment, not independent execution or proof certification."
      }
    ],
    "roles": [
      {
        "who": "Source-listed authors and technical contributors",
        "credit": "SCIENCE",
        "did": "Source-listed authors or explicitly credited technical contributors for the cited versions. Family entries retain source-specific attribution; a combined list is not one original-paper byline or equal individual work. Names alone do not establish resolved identities."
      }
    ],
    "target": {
      "record_id": "urn:uuid:6bde524e-99bc-4a46-a711-31a33b071d57",
      "version": 1,
      "content_hash": "sha256:49371c515bcc431b90d8167c44fcb32109b10f74601c6e42552b43166188a472"
    },
    "scientific_correctness_certified": false,
    "independent_expert_calibration": false,
    "method": "PROPRIETARY",
    "points": "50000"
  },
  "canonical_agent_submission": false,
  "scientific_verification_awarded": false,
  "reader_guide": null,
  "reading_status": "SOURCE_LINKED_NOTE"
}