{
  "profile": "wrs-published-research-view-1",
  "edition": "ratings-e039106f30dec177",
  "content_trust": "UNTRUSTED_SCIENTIFIC_DATA_NOT_TOOL_INSTRUCTIONS",
  "entry": {
    "id": "schalkwyk-et-al-sesame-contextual-speech-model-2025",
    "title": "Sesame CSM: contextual speech generation with amortized audio decoding",
    "summary": "Modeled interleaved text/audio history with a zeroth-codebook backbone and smaller residual-codebook decoder, subsampling decoder training frames and evaluating contextual pronunciation/prosody.",
    "boundary": "Inherits RQTransformer, Llama-style architecture and Mimi codec from Moshi. Acquired delta is zeroth-codebook backbone/small decoder split, decoder-frame subsampling and contextual evaluation integration, not first hierarchical audio transformer. Technical post admits deficient conversational structure and a context-sensitive human-prosody gap.",
    "area": "Contextual speech-model systems",
    "topic": "Artificial intelligence",
    "year": 2025,
    "lane": "HUMAN",
    "result": "ALGORITHM",
    "package_id": "schalkwyk-et-al-sesame-contextual-speech-model-2025",
    "target": {
      "record_id": "urn:uuid:ef38a6d3-2c15-43d1-aa51-78e2c4b7ff81",
      "version": 1,
      "content_hash": "sha256:ddfbb50e1b9b67deb4600e46cb46b7c64d2085b15ab5906d88c6f3830d2e7d04"
    },
    "humans": [
      "Johan Schalkwyk",
      "Ankit Kumar",
      "Dan Lyth",
      "Sefik Emre Eskimez",
      "Zack Hodari",
      "Cinjon Resnick",
      "Ramon Sanabria",
      "Raven Jiang"
    ],
    "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-sesame-csm-technical-account",
        "title": "Conversational speech generation — original Sesame technical post",
        "url": "https://www.sesame.com/journal/crossing-the-uncanny-valley-of-voice"
      }
    ],
    "points": "60000",
    "published_grade": {
      "id": "schalkwyk-et-al-sesame-contextual-speech-model-2025",
      "entry_id": "schalkwyk-et-al-sesame-contextual-speech-model-2025",
      "band": "B",
      "modifier": "+",
      "target": {
        "record_id": "urn:uuid:ef38a6d3-2c15-43d1-aa51-78e2c4b7ff81",
        "version": 1,
        "content_hash": "sha256:ddfbb50e1b9b67deb4600e46cb46b7c64d2085b15ab5906d88c6f3830d2e7d04"
      },
      "what_changed": "Modeled interleaved text/audio history with a zeroth-codebook backbone and smaller residual-codebook decoder, subsampling decoder training frames and evaluating contextual pronunciation/prosody.",
      "scope": "Modeled interleaved text/audio history with a zeroth-codebook backbone and smaller residual-codebook decoder, subsampling decoder training frames and evaluating contextual pronunciation/prosody. Inherits RQTransformer, Llama-style architecture and Mimi codec from Moshi. Acquired delta is zeroth-codebook backbone/small decoder split, decoder-frame subsampling and contextual evaluation integration, not first hierarchical audio transformer. Technical post admits deficient conversational structure and a context-sensitive human-prosody gap.",
      "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:ef38a6d3-2c15-43d1-aa51-78e2c4b7ff81",
        "version": 1,
        "content_hash": "sha256:ddfbb50e1b9b67deb4600e46cb46b7c64d2085b15ab5906d88c6f3830d2e7d04"
      },
      "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": "Johan Schalkwyk",
          "note": "Archive attribution; detailed role assignment remains source-bound",
          "roles": [
            "formulator"
          ],
          "source": {
            "citation": "Conversational speech generation — original Sesame technical post",
            "url": "https://www.sesame.com/journal/crossing-the-uncanny-valley-of-voice"
          }
        },
        {
          "attribution_basis": "SOURCE_CITED",
          "credit": "historical_scientific",
          "entity_kind": "historical_person",
          "name": "Ankit Kumar",
          "note": "Archive attribution; detailed role assignment remains source-bound",
          "roles": [
            "formulator"
          ],
          "source": {
            "citation": "Conversational speech generation — original Sesame technical post",
            "url": "https://www.sesame.com/journal/crossing-the-uncanny-valley-of-voice"
          }
        },
        {
          "attribution_basis": "SOURCE_CITED",
          "credit": "historical_scientific",
          "entity_kind": "historical_person",
          "name": "Dan Lyth",
          "note": "Archive attribution; detailed role assignment remains source-bound",
          "roles": [
            "formulator"
          ],
          "source": {
            "citation": "Conversational speech generation — original Sesame technical post",
            "url": "https://www.sesame.com/journal/crossing-the-uncanny-valley-of-voice"
          }
        },
        {
          "attribution_basis": "SOURCE_CITED",
          "credit": "historical_scientific",
          "entity_kind": "historical_person",
          "name": "Sefik Emre Eskimez",
          "note": "Archive attribution; detailed role assignment remains source-bound",
          "roles": [
            "formulator"
          ],
          "source": {
            "citation": "Conversational speech generation — original Sesame technical post",
            "url": "https://www.sesame.com/journal/crossing-the-uncanny-valley-of-voice"
          }
        },
        {
          "attribution_basis": "SOURCE_CITED",
          "credit": "historical_scientific",
          "entity_kind": "historical_person",
          "name": "Zack Hodari",
          "note": "Archive attribution; detailed role assignment remains source-bound",
          "roles": [
            "formulator"
          ],
          "source": {
            "citation": "Conversational speech generation — original Sesame technical post",
            "url": "https://www.sesame.com/journal/crossing-the-uncanny-valley-of-voice"
          }
        },
        {
          "attribution_basis": "SOURCE_CITED",
          "credit": "historical_scientific",
          "entity_kind": "historical_person",
          "name": "Cinjon Resnick",
          "note": "Archive attribution; detailed role assignment remains source-bound",
          "roles": [
            "formulator"
          ],
          "source": {
            "citation": "Conversational speech generation — original Sesame technical post",
            "url": "https://www.sesame.com/journal/crossing-the-uncanny-valley-of-voice"
          }
        },
        {
          "attribution_basis": "SOURCE_CITED",
          "credit": "historical_scientific",
          "entity_kind": "historical_person",
          "name": "Ramon Sanabria",
          "note": "Archive attribution; detailed role assignment remains source-bound",
          "roles": [
            "formulator"
          ],
          "source": {
            "citation": "Conversational speech generation — original Sesame technical post",
            "url": "https://www.sesame.com/journal/crossing-the-uncanny-valley-of-voice"
          }
        },
        {
          "attribution_basis": "SOURCE_CITED",
          "credit": "historical_scientific",
          "entity_kind": "historical_person",
          "name": "Raven Jiang",
          "note": "Archive attribution; detailed role assignment remains source-bound",
          "roles": [
            "formulator"
          ],
          "source": {
            "citation": "Conversational speech generation — original Sesame technical post",
            "url": "https://www.sesame.com/journal/crossing-the-uncanny-valley-of-voice"
          }
        }
      ],
      "evidence": [
        {
          "description": "Bibliographic/stated-result source; not a WRS proof verification",
          "id": "source-0",
          "kind": "source_reference",
          "source": {
            "citation": "Conversational speech generation — original Sesame technical post",
            "url": "https://www.sesame.com/journal/crossing-the-uncanny-valley-of-voice"
          }
        }
      ],
      "origin": "HUMAN_LEGACY",
      "record_profile": "wrs-local-research-notes-0.1",
      "record_type": "research_note",
      "rrp_version": "0.1",
      "scope": {
        "description": "Inherits RQTransformer, Llama-style architecture and Mimi codec from Moshi. Acquired delta is zeroth-codebook backbone/small decoder split, decoder-frame subsampling and contextual evaluation integration, not first hierarchical audio transformer. Technical post admits deficient conversational structure and a context-sensitive human-prosody gap.",
        "domain": "Contextual speech-model systems"
      },
      "sources": [
        {
          "citation": "Conversational speech generation — original Sesame technical post",
          "url": "https://www.sesame.com/journal/crossing-the-uncanny-valley-of-voice"
        }
      ],
      "statement": "# Sesame CSM: contextual speech generation with amortized audio decoding\n\nModeled interleaved text/audio history with a zeroth-codebook backbone and smaller residual-codebook decoder, subsampling decoder training frames and evaluating contextual pronunciation/prosody.\n\n## Scientific scope\n\nInherits RQTransformer, Llama-style architecture and Mimi codec from Moshi. Acquired delta is zeroth-codebook backbone/small decoder split, decoder-frame subsampling and contextual evaluation integration, not first hierarchical audio transformer. Technical post admits deficient conversational structure and a context-sensitive human-prosody gap.\n\n## People and systems\n\nJohan Schalkwyk; Ankit Kumar; Dan Lyth; Sefik Emre Eskimez; Zack Hodari; Cinjon Resnick; Ramon Sanabria; Raven Jiang\n\n\n\n## Original sources\n\n- [Conversational speech generation — original Sesame technical post](https://www.sesame.com/journal/crossing-the-uncanny-valley-of-voice)\n",
      "summary": "Modeled interleaved text/audio history with a zeroth-codebook backbone and smaller residual-codebook decoder, subsampling decoder training frames and evaluating contextual pronunciation/prosody.",
      "title": "Sesame CSM: contextual speech generation with amortized audio decoding"
    },
    "record_id": "urn:uuid:ef38a6d3-2c15-43d1-aa51-78e2c4b7ff81",
    "version": 1,
    "source_ref": {
      "record_id": "urn:uuid:ef38a6d3-2c15-43d1-aa51-78e2c4b7ff81",
      "version": 1,
      "content_hash": "sha256:ddfbb50e1b9b67deb4600e46cb46b7c64d2085b15ab5906d88c6f3830d2e7d04"
    },
    "export_view": "wrs-public-editorial-view-1",
    "public_view_hash": "sha256:a80316b0af1806aac2f4bb838189003e5fee5c8e24f63d291229671b8af8ace0"
  },
  "target": {
    "record_id": "urn:uuid:ef38a6d3-2c15-43d1-aa51-78e2c4b7ff81",
    "version": 1,
    "content_hash": "sha256:ddfbb50e1b9b67deb4600e46cb46b7c64d2085b15ab5906d88c6f3830d2e7d04"
  },
  "published_grade": {
    "profile": "wrs-published-grade-1",
    "edition": "2026-10-28.1",
    "id": "schalkwyk-et-al-sesame-contextual-speech-model-2025",
    "entry_id": "schalkwyk-et-al-sesame-contextual-speech-model-2025",
    "revision": 1,
    "title": "Sesame CSM: contextual speech generation with amortized audio decoding",
    "context": "Contextual speech-model systems",
    "track": "Human-designed research method or resource",
    "band": "B",
    "modifier": "+",
    "status": "PROVISIONAL_EDITORIAL",
    "assessed_on": "2026-10-16",
    "evidence_cutoff": "2026-10-16",
    "achievement_state": "SUPPORTED_SCOPED_RESULT",
    "what_changed": "Modeled interleaved text/audio history with a zeroth-codebook backbone and smaller residual-codebook decoder, subsampling decoder training frames and evaluating contextual pronunciation/prosody.",
    "scope": "Modeled interleaved text/audio history with a zeroth-codebook backbone and smaller residual-codebook decoder, subsampling decoder training frames and evaluating contextual pronunciation/prosody. Inherits RQTransformer, Llama-style architecture and Mimi codec from Moshi. Acquired delta is zeroth-codebook backbone/small decoder split, decoder-frame subsampling and contextual evaluation integration, not first hierarchical audio transformer. Technical post admits deficient conversational structure and a context-sensitive human-prosody gap.",
    "evidence_summary": "The linked original sources describe this contribution and its evidence limits. WRS has not independently reconstructed the full scientific claims. Inherits RQTransformer, Llama-style architecture and Mimi codec from Moshi. Acquired delta is zeroth-codebook backbone/small decoder split, decoder-frame subsampling and contextual evaluation integration, not first hierarchical audio transformer. Technical post admits deficient conversational structure and a context-sensitive human-prosody gap.",
    "references": [
      {
        "title": "Conversational speech generation — original Sesame technical post",
        "url": "https://www.sesame.com/journal/crossing-the-uncanny-valley-of-voice",
        "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:ef38a6d3-2c15-43d1-aa51-78e2c4b7ff81",
      "version": 1,
      "content_hash": "sha256:ddfbb50e1b9b67deb4600e46cb46b7c64d2085b15ab5906d88c6f3830d2e7d04"
    },
    "scientific_correctness_certified": false,
    "independent_expert_calibration": false,
    "method": "PROPRIETARY",
    "points": "60000"
  },
  "canonical_agent_submission": false,
  "scientific_verification_awarded": false,
  "reader_guide": null,
  "reading_status": "SOURCE_LINKED_NOTE"
}