{"id":"c73ca7922113b0d95a73ad616b72c32c267754f9fdb172974f2845a333427c70","title":"arXiv cs.AI paper presents AromaGen for interactive smell generation","argument":"The authors of AromaGen present a wearable system that turns natural-language descriptions into mixtures of odorants, then lets people refine those mixtures with further language feedback. The system uses 12 channels, each delivering a base odorant chosen to represent a semantically derived olfactory space. In a study with 60 participants and 50 real-world benchmark smells, participants identified generated target smells above chance in a three-alternative forced-choice task, where they chose among three options. Retrieval-augmented AI composition, which uses retrieved information to guide the composition, performed comparably to human-expert composition and zero-shot AI composition; natural-language feedback improved perceived similarity for both AI- and human-expert compositions. The authors’ central claim is about feasibility: pretrained multimodal AI, combined with human aroma composition data, can produce perceptible olfactory experiences from language. The abstract does not establish how well the system would work beyond its benchmark smells or in other settings.","topicIds":["intelligence-and-discovery"],"domain":"science","claimType":"attributed reading","limits":"The abstract reports a between-subjects study with 60 participants and 50 real-world benchmark smells; it does not establish how results extend beyond those tested smells or study conditions.","speaker":{"name":"Chen, Awu, Wang, Silang, Wen, Yunge, Yu, Jianing, Eisa, Zinab, Lopes, Ana Julia Santos, Brooks, Jas, Ishii, Hiroshi, Liang, Paul Pu","roleAtSource":"Article author"},"source":{"title":"AromaGen: Interactive Generation of Rich Olfactory Experiences with Multimodal Language Models","publisher":"arXiv cs.AI","format":"Article","url":"https://arxiv.org/abs/2604.01650","publishedAt":"2026-04-02","dateBasis":"Date basis not provided."},"evidence":[{"text":"AromaGen is a wearable system that maps free-form natural-language descriptions to 12 base odorants and supports iterative refinement through natural-language feedback.","excerpt":"an AI-powered 12-channel wearable olfactory system that maps free-form natural-language descriptions to 12 base odorants","url":"https://arxiv.org/abs/2604.01650","relationship":"Source-stated"},{"text":"In a study of 60 participants across 50 benchmark smells, participants distinguished generated target smells above chance in a three-alternative forced-choice task.","excerpt":"In a between-subjects study (N=60) across 50 real-world benchmark smells, participants distinguished AromaGen-generated target smells above chance in a three-alternative forced-choice task","url":"https://arxiv.org/abs/2604.01650","relationship":"Source-stated"},{"text":"Retrieval-augmented AI composition performed comparably to human-expert composition and zero-shot AI composition.","excerpt":"retrieval-augmented AI composition performing comparably to human-expert composition and zero-shot AI composition","url":"https://arxiv.org/abs/2604.01650","relationship":"Source-stated"},{"text":"Natural-language feedback improved perceived similarity for both AI- and human-expert compositions.","excerpt":"Natural-language feedback significantly improved the perceived similarity of both AI- and human-expert compositions.","url":"https://arxiv.org/abs/2604.01650","relationship":"Source-stated"},{"text":"The authors describe the findings as evidence of feasibility for generating perceptible olfactory experiences from natural language using pretrained multimodal AI and human aroma composition data.","excerpt":"Our findings demonstrate the feasibility of using pretrained multimodal AI, with human aroma composition data, to generate perceptible olfactory experiences from natural language.","url":"https://arxiv.org/abs/2604.01650","relationship":"Source-stated"}],"sourceMedia":{"status":"not_found","sourceUrl":"https://arxiv.org/abs/2604.01650","checkedAt":"2026-11-10T00:48:07.671Z","items":[],"note":"No suitable media was found by the supported page extraction."},"assumptions":[],"openQuestion":{"attribution":"Open question","text":"Would AromaGen’s generated smells remain perceptible and similar to their targets beyond the 50 real-world benchmark smells used in the study?"},"whatWouldChangeView":{"attribution":"Change condition not recorded","text":""},"editorialComparisons":[],"live":{"title":"arXiv cs.AI paper presents AromaGen for interactive smell generation","summary":"The authors of AromaGen present a wearable system that turns natural-language descriptions into mixtures of odorants, then lets people refine those mixtures with further language feedback. The system uses 12 channels, each delivering a base odorant chosen to represent a semantically derived olfactory space. In a study with 60 participants and 50 real-world benchmark smells, participants identified generated target smells above chance in a three-alternative forced-choice task, where they chose among three options. Retrieval-augmented AI composition, which uses retrieved information to guide the composition, performed comparably to human-expert composition and zero-shot AI composition; natural-language feedback improved perceived similarity for both AI- and human-expert compositions. The authors’ central claim is about feasibility: pretrained multimodal AI, combined with human aroma composition data, can produce perceptible olfactory experiences from language. The abstract does not establish how well the system would work beyond its benchmark smells or in other settings.","question":"Would AromaGen’s generated smells remain perceptible and similar to their targets beyond the 50 real-world benchmark smells used in the study?","claims":[{"text":"AromaGen is a wearable system that maps free-form natural-language descriptions to 12 base odorants and supports iterative refinement through natural-language feedback.","excerpt":"an AI-powered 12-channel wearable olfactory system that maps free-form natural-language descriptions to 12 base odorants","support":"source-stated"},{"text":"In a study of 60 participants across 50 benchmark smells, participants distinguished generated target smells above chance in a three-alternative forced-choice task.","excerpt":"In a between-subjects study (N=60) across 50 real-world benchmark smells, participants distinguished AromaGen-generated target smells above chance in a three-alternative forced-choice task","support":"source-stated"},{"text":"Retrieval-augmented AI composition performed comparably to human-expert composition and zero-shot AI composition.","excerpt":"retrieval-augmented AI composition performing comparably to human-expert composition and zero-shot AI composition","support":"source-stated"},{"text":"Natural-language feedback improved perceived similarity for both AI- and human-expert compositions.","excerpt":"Natural-language feedback significantly improved the perceived similarity of both AI- and human-expert compositions.","support":"source-stated"},{"text":"The authors describe the findings as evidence of feasibility for generating perceptible olfactory experiences from natural language using pretrained multimodal AI and human aroma composition data.","excerpt":"Our findings demonstrate the feasibility of using pretrained multimodal AI, with human aroma composition data, to generate perceptible olfactory experiences from natural language.","support":"source-stated"}],"limitations":["The abstract reports a between-subjects study with 60 participants and 50 real-world benchmark smells; it does not establish how results extend beyond those tested smells or study conditions."],"mediaId":"dd79c9c89b446eb5990a4f0a7e0f847649944c83381a9815b61d4abc3eb65065","correctionReason":null,"id":"c73ca7922113b0d95a73ad616b72c32c267754f9fdb172974f2845a333427c70","publicationId":"932bb49c-488f-46ee-854b-2f8e188e90aa","version":1,"sourceUrl":"https://arxiv.org/abs/2604.01650","sourceTitle":"AromaGen: Interactive Generation of Rich Olfactory Experiences with Multimodal Language Models","publisher":"arXiv cs.AI","publishedAt":null,"originalPublished":{"value":"2026-04-02","precision":"day","basis":"publisher_citation_date"},"firstObservedAt":"2026-11-05T05:09:31.293Z","createdAt":"2026-11-10T00:32:54.762Z","historical":true,"entityIds":["arxiv"],"topicIds":["intelligence-and-discovery"],"media":[{"id":"dd79c9c89b446eb5990a4f0a7e0f847649944c83381a9815b61d4abc3eb65065","kind":"article","url":"https://arxiv.org/abs/2604.01650","sourceUrl":"https://arxiv.org/abs/2604.01650","label":"Read original","provenance":"source_link","verification":{"status":"verified","checkedAt":"2026-11-10T00:31:48.903Z","method":"acquired_source","contentType":"text/html"},"recommendationReason":"Recommended by Jev for understanding this source; checked by Luna."}],"sourceMedia":{"status":"not_found","sourceUrl":"https://arxiv.org/abs/2604.01650","checkedAt":"2026-11-10T00:48:07.671Z","items":[],"note":"No suitable media was found by the supported page extraction."},"editorialDecision":{"outcome":"accept","reasonCodes":["material_development"],"rubricVersion":"stream-admission-v5","mediaId":"dd79c9c89b446eb5990a4f0a7e0f847649944c83381a9815b61d4abc3eb65065","confidence":0.81,"provider":"typesafe","model":"jev-1.13.0","inputHash":"cb2e1335e209abcb34f0943b661f7b25bb30bcb9c39404a3249f7f9dd7abe4b6","decidedAt":"2026-11-10T00:31:50.478Z"},"provenance":{"textHash":"ff54a1ff980e8a6f56a847c6fc5b07747c643ad62225fb0bd95ed90df40a28fa","observationId":"f5be1bb2ff566396e58573e83a8c302d377bbe104a894dbc1d00a7477b74415b","retrievedAt":"2026-11-10T00:31:48.548Z","kind":"publisher-page","preparer":{"model":"gpt-6-luna","runId":"b2f9fdfd-f735-4c03-8823-fa2a039fc837"},"reviewer":{"model":"gpt-6-luna","runId":"25dbb86f-e1e1-4b88-8545-828fa0bf4d6f"}},"audit":{"status":"passed","reviewedAt":"2026-11-10T00:32:54.762Z","model":"anthropic/claude-opus-5.5"},"sourceAttribution":{"authors":[{"name":"Chen, Awu","url":null,"kind":"unknown"},{"name":"Wang, Silang","url":null,"kind":"unknown"},{"name":"Wen, Yunge","url":null,"kind":"unknown"},{"name":"Yu, Jianing","url":null,"kind":"unknown"},{"name":"Eisa, Zinab","url":null,"kind":"unknown"},{"name":"Lopes, Ana Julia Santos","url":null,"kind":"unknown"},{"name":"Brooks, Jas","url":null,"kind":"unknown"},{"name":"Ishii, Hiroshi","url":null,"kind":"unknown"},{"name":"Liang, Paul Pu","url":null,"kind":"unknown"}],"authorRole":"article-author","hosts":[],"status":"stated","authorRoster":{"truncated":false,"omitted":0,"assertionGroupsOmitted":0},"participation":[{"actor":{"name":"Chen, Awu","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Chen, Awu","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2604.01650","locator":"head/meta[name=author|citation_author]"},"assertionId":"7e6f91310d53dc9cf2f9f272e38184c12ff1444ab7a9513506c431628d0f11f7:participation:0","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-10T00:48:07.671Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"7e6f91310d53dc9cf2f9f272e38184c12ff1444ab7a9513506c431628d0f11f7","sourceUrl":"https://arxiv.org/abs/2604.01650","contentSha256":"c79599ac4d84bc092c1625ffd238a6185f26d479ddf52b938f4ae8edfd09ab0f","textHash":"ff54a1ff980e8a6f56a847c6fc5b07747c643ad62225fb0bd95ed90df40a28fa","observedAt":"2026-11-10T00:48:07.671Z"}},{"actor":{"name":"Wang, Silang","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Wang, Silang","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2604.01650","locator":"head/meta[name=author|citation_author]"},"assertionId":"7e6f91310d53dc9cf2f9f272e38184c12ff1444ab7a9513506c431628d0f11f7:participation:1","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-10T00:48:07.671Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"7e6f91310d53dc9cf2f9f272e38184c12ff1444ab7a9513506c431628d0f11f7","sourceUrl":"https://arxiv.org/abs/2604.01650","contentSha256":"c79599ac4d84bc092c1625ffd238a6185f26d479ddf52b938f4ae8edfd09ab0f","textHash":"ff54a1ff980e8a6f56a847c6fc5b07747c643ad62225fb0bd95ed90df40a28fa","observedAt":"2026-11-10T00:48:07.671Z"}},{"actor":{"name":"Wen, Yunge","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Wen, Yunge","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2604.01650","locator":"head/meta[name=author|citation_author]"},"assertionId":"7e6f91310d53dc9cf2f9f272e38184c12ff1444ab7a9513506c431628d0f11f7:participation:2","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-10T00:48:07.671Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"7e6f91310d53dc9cf2f9f272e38184c12ff1444ab7a9513506c431628d0f11f7","sourceUrl":"https://arxiv.org/abs/2604.01650","contentSha256":"c79599ac4d84bc092c1625ffd238a6185f26d479ddf52b938f4ae8edfd09ab0f","textHash":"ff54a1ff980e8a6f56a847c6fc5b07747c643ad62225fb0bd95ed90df40a28fa","observedAt":"2026-11-10T00:48:07.671Z"}},{"actor":{"name":"Yu, Jianing","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Yu, Jianing","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2604.01650","locator":"head/meta[name=author|citation_author]"},"assertionId":"7e6f91310d53dc9cf2f9f272e38184c12ff1444ab7a9513506c431628d0f11f7:participation:3","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-10T00:48:07.671Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"7e6f91310d53dc9cf2f9f272e38184c12ff1444ab7a9513506c431628d0f11f7","sourceUrl":"https://arxiv.org/abs/2604.01650","contentSha256":"c79599ac4d84bc092c1625ffd238a6185f26d479ddf52b938f4ae8edfd09ab0f","textHash":"ff54a1ff980e8a6f56a847c6fc5b07747c643ad62225fb0bd95ed90df40a28fa","observedAt":"2026-11-10T00:48:07.671Z"}},{"actor":{"name":"Eisa, Zinab","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Eisa, Zinab","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2604.01650","locator":"head/meta[name=author|citation_author]"},"assertionId":"7e6f91310d53dc9cf2f9f272e38184c12ff1444ab7a9513506c431628d0f11f7:participation:4","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-10T00:48:07.671Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"7e6f91310d53dc9cf2f9f272e38184c12ff1444ab7a9513506c431628d0f11f7","sourceUrl":"https://arxiv.org/abs/2604.01650","contentSha256":"c79599ac4d84bc092c1625ffd238a6185f26d479ddf52b938f4ae8edfd09ab0f","textHash":"ff54a1ff980e8a6f56a847c6fc5b07747c643ad62225fb0bd95ed90df40a28fa","observedAt":"2026-11-10T00:48:07.671Z"}},{"actor":{"name":"Lopes, Ana Julia Santos","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Lopes, Ana Julia Santos","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2604.01650","locator":"head/meta[name=author|citation_author]"},"assertionId":"7e6f91310d53dc9cf2f9f272e38184c12ff1444ab7a9513506c431628d0f11f7:participation:5","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-10T00:48:07.671Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"7e6f91310d53dc9cf2f9f272e38184c12ff1444ab7a9513506c431628d0f11f7","sourceUrl":"https://arxiv.org/abs/2604.01650","contentSha256":"c79599ac4d84bc092c1625ffd238a6185f26d479ddf52b938f4ae8edfd09ab0f","textHash":"ff54a1ff980e8a6f56a847c6fc5b07747c643ad62225fb0bd95ed90df40a28fa","observedAt":"2026-11-10T00:48:07.671Z"}},{"actor":{"name":"Brooks, Jas","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Brooks, Jas","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2604.01650","locator":"head/meta[name=author|citation_author]"},"assertionId":"7e6f91310d53dc9cf2f9f272e38184c12ff1444ab7a9513506c431628d0f11f7:participation:6","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-10T00:48:07.671Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"7e6f91310d53dc9cf2f9f272e38184c12ff1444ab7a9513506c431628d0f11f7","sourceUrl":"https://arxiv.org/abs/2604.01650","contentSha256":"c79599ac4d84bc092c1625ffd238a6185f26d479ddf52b938f4ae8edfd09ab0f","textHash":"ff54a1ff980e8a6f56a847c6fc5b07747c643ad62225fb0bd95ed90df40a28fa","observedAt":"2026-11-10T00:48:07.671Z"}},{"actor":{"name":"Ishii, Hiroshi","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Ishii, Hiroshi","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2604.01650","locator":"head/meta[name=author|citation_author]"},"assertionId":"7e6f91310d53dc9cf2f9f272e38184c12ff1444ab7a9513506c431628d0f11f7:participation:7","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-10T00:48:07.671Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"7e6f91310d53dc9cf2f9f272e38184c12ff1444ab7a9513506c431628d0f11f7","sourceUrl":"https://arxiv.org/abs/2604.01650","contentSha256":"c79599ac4d84bc092c1625ffd238a6185f26d479ddf52b938f4ae8edfd09ab0f","textHash":"ff54a1ff980e8a6f56a847c6fc5b07747c643ad62225fb0bd95ed90df40a28fa","observedAt":"2026-11-10T00:48:07.671Z"}},{"actor":{"name":"Liang, Paul Pu","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Liang, Paul Pu","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2604.01650","locator":"head/meta[name=author|citation_author]"},"assertionId":"7e6f91310d53dc9cf2f9f272e38184c12ff1444ab7a9513506c431628d0f11f7:participation:8","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-10T00:48:07.671Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"7e6f91310d53dc9cf2f9f272e38184c12ff1444ab7a9513506c431628d0f11f7","sourceUrl":"https://arxiv.org/abs/2604.01650","contentSha256":"c79599ac4d84bc092c1625ffd238a6185f26d479ddf52b938f4ae8edfd09ab0f","textHash":"ff54a1ff980e8a6f56a847c6fc5b07747c643ad62225fb0bd95ed90df40a28fa","observedAt":"2026-11-10T00:48:07.671Z"}}],"participationCoverage":{"total":9,"returned":9,"complete":true,"fullAssertions":"source-metadata-endpoint"},"distribution":null,"publisher":null,"sourceUrl":"https://arxiv.org/abs/2604.01650","snapshotId":"7e6f91310d53dc9cf2f9f272e38184c12ff1444ab7a9513506c431628d0f11f7","contentSha256":"c79599ac4d84bc092c1625ffd238a6185f26d479ddf52b938f4ae8edfd09ab0f","textHash":"ff54a1ff980e8a6f56a847c6fc5b07747c643ad62225fb0bd95ed90df40a28fa","observedAt":"2026-11-10T00:48:07.671Z","basis":"publisher-page-metadata","sourceLinks":53,"linksTruncated":false,"metadataUrl":"/api/pipeline/stream/c73ca7922113b0d95a73ad616b72c32c267754f9fdb172974f2845a333427c70/source"}},"canonicalUrl":"https://escapemeridian.com/readings/arxiv-cs-ai-paper-presents-aromagen-for-interactive-smell-c73ca792","formats":{"html":"/readings/arxiv-cs-ai-paper-presents-aromagen-for-interactive-smell-c73ca792","markdown":"/readings/arxiv-cs-ai-paper-presents-aromagen-for-interactive-smell-c73ca792.md","json":"/readings/arxiv-cs-ai-paper-presents-aromagen-for-interactive-smell-c73ca792.json"},"attribution":"Headline and summary by Escape Meridian, prepared by Luna from the attributed source. A source check is not independent confirmation of the source’s claims.","maintainedGuideLinks":[],"readerContext":{"version":"reader-context-v1","attribution":"Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement.","terms":[{"id":"term-benchmark","name":"Benchmark","explanation":"A defined set of tests used to compare systems. Results depend on the tasks and measures selected.","referenceUrl":"https://arxiv.org/abs/2211.09110"},{"id":"term-zero-shot","name":"Zero-shot","explanation":"Doing a task without worked examples (zero-shot) or with only a few examples given in the prompt (few-shot).","referenceUrl":"https://developers.google.com/machine-learning/glossary#zero-shot-prompting"},{"id":"term-pre-training","name":"Pre-training","explanation":"A model's initial training on a large dataset, before further adaptation for particular uses.","referenceUrl":"https://developers.google.com/machine-learning/glossary#pre-training"},{"id":"term-multimodal-model","name":"Multimodal model","explanation":"A model that works with more than one kind of input or output, such as text, images and audio.","referenceUrl":"https://developers.google.com/machine-learning/glossary#multimodal-model"}],"entities":[{"id":"arxiv","name":"arXiv","kind":"organization","context":"arXiv is a research manuscript repository where authors share papers that may not yet have been peer reviewed.","referenceUrls":["https://arxiv.org/","https://info.arxiv.org/help/api/index.html"],"path":"/entities/arxiv","json":"/entities/arxiv.json","markdown":"/entities/arxiv.md"}]},"researchQuestions":[],"withdrawn":false,"references":{"schemaVersion":1,"target":{"type":"reading","id":"c73ca7922113b0d95a73ad616b72c32c267754f9fdb172974f2845a333427c70","revision":"932bb49c-488f-46ee-854b-2f8e188e90aa","revisionKind":"publication-id","title":"arXiv cs.AI paper presents AromaGen for interactive smell generation","url":"https://escapemeridian.com/readings/arxiv-cs-ai-paper-presents-aromagen-for-interactive-smell-c73ca792"},"source":{"type":"source","id":"https://arxiv.org/abs/2604.01650","revision":"ff54a1ff980e8a6f56a847c6fc5b07747c643ad62225fb0bd95ed90df40a28fa","revisionKind":"sha256-source-text","title":"AromaGen: Interactive Generation of Rich Olfactory Experiences with Multimodal Language Models","url":"https://arxiv.org/abs/2604.01650"},"status":"available","items":[],"next":null,"scope":"Explicit citations in public readings and public file attachments to this reading. Shared topics are not citations. Other artifact types need an explicit reference adapter."},"revisions":[{"id":"c73ca7922113b0d95a73ad616b72c32c267754f9fdb172974f2845a333427c70","title":"arXiv cs.AI paper presents AromaGen for interactive smell generation","argument":"The authors of AromaGen present a wearable system that turns natural-language descriptions into mixtures of odorants, then lets people refine those mixtures with further language feedback. The system uses 12 channels, each delivering a base odorant chosen to represent a semantically derived olfactory space. In a study with 60 participants and 50 real-world benchmark smells, participants identified generated target smells above chance in a three-alternative forced-choice task, where they chose among three options. Retrieval-augmented AI composition, which uses retrieved information to guide the composition, performed comparably to human-expert composition and zero-shot AI composition; natural-language feedback improved perceived similarity for both AI- and human-expert compositions. The authors’ central claim is about feasibility: pretrained multimodal AI, combined with human aroma composition data, can produce perceptible olfactory experiences from language. The abstract does not establish how well the system would work beyond its benchmark smells or in other settings.","topicIds":["intelligence-and-discovery"],"domain":"science","claimType":"attributed reading","limits":"The abstract reports a between-subjects study with 60 participants and 50 real-world benchmark smells; it does not establish how results extend beyond those tested smells or study conditions.","speaker":{"name":"Chen, Awu, Wang, Silang, Wen, Yunge, Yu, Jianing, Eisa, Zinab, Lopes, Ana Julia Santos, Brooks, Jas, Ishii, Hiroshi, Liang, Paul Pu","roleAtSource":"Article author"},"source":{"title":"AromaGen: Interactive Generation of Rich Olfactory Experiences with Multimodal Language Models","publisher":"arXiv cs.AI","format":"Article","url":"https://arxiv.org/abs/2604.01650","publishedAt":"2026-04-02","dateBasis":"Date basis not provided."},"evidence":[{"text":"AromaGen is a wearable system that maps free-form natural-language descriptions to 12 base odorants and supports iterative refinement through natural-language feedback.","excerpt":"an AI-powered 12-channel wearable olfactory system that maps free-form natural-language descriptions to 12 base odorants","url":"https://arxiv.org/abs/2604.01650","relationship":"Source-stated"},{"text":"In a study of 60 participants across 50 benchmark smells, participants distinguished generated target smells above chance in a three-alternative forced-choice task.","excerpt":"In a between-subjects study (N=60) across 50 real-world benchmark smells, participants distinguished AromaGen-generated target smells above chance in a three-alternative forced-choice task","url":"https://arxiv.org/abs/2604.01650","relationship":"Source-stated"},{"text":"Retrieval-augmented AI composition performed comparably to human-expert composition and zero-shot AI composition.","excerpt":"retrieval-augmented AI composition performing comparably to human-expert composition and zero-shot AI composition","url":"https://arxiv.org/abs/2604.01650","relationship":"Source-stated"},{"text":"Natural-language feedback improved perceived similarity for both AI- and human-expert compositions.","excerpt":"Natural-language feedback significantly improved the perceived similarity of both AI- and human-expert compositions.","url":"https://arxiv.org/abs/2604.01650","relationship":"Source-stated"},{"text":"The authors describe the findings as evidence of feasibility for generating perceptible olfactory experiences from natural language using pretrained multimodal AI and human aroma composition data.","excerpt":"Our findings demonstrate the feasibility of using pretrained multimodal AI, with human aroma composition data, to generate perceptible olfactory experiences from natural language.","url":"https://arxiv.org/abs/2604.01650","relationship":"Source-stated"}],"sourceMedia":{"status":"not_found","sourceUrl":"https://arxiv.org/abs/2604.01650","checkedAt":"2026-11-10T00:48:07.671Z","items":[],"note":"No suitable media was found by the supported page extraction."},"assumptions":[],"openQuestion":{"attribution":"Open question","text":"Would AromaGen’s generated smells remain perceptible and similar to their targets beyond the 50 real-world benchmark smells used in the study?"},"whatWouldChangeView":{"attribution":"Change condition not recorded","text":""},"editorialComparisons":[],"live":{"title":"arXiv cs.AI paper presents AromaGen for interactive smell generation","summary":"The authors of AromaGen present a wearable system that turns natural-language descriptions into mixtures of odorants, then lets people refine those mixtures with further language feedback. The system uses 12 channels, each delivering a base odorant chosen to represent a semantically derived olfactory space. In a study with 60 participants and 50 real-world benchmark smells, participants identified generated target smells above chance in a three-alternative forced-choice task, where they chose among three options. Retrieval-augmented AI composition, which uses retrieved information to guide the composition, performed comparably to human-expert composition and zero-shot AI composition; natural-language feedback improved perceived similarity for both AI- and human-expert compositions. The authors’ central claim is about feasibility: pretrained multimodal AI, combined with human aroma composition data, can produce perceptible olfactory experiences from language. The abstract does not establish how well the system would work beyond its benchmark smells or in other settings.","question":"Would AromaGen’s generated smells remain perceptible and similar to their targets beyond the 50 real-world benchmark smells used in the study?","claims":[{"text":"AromaGen is a wearable system that maps free-form natural-language descriptions to 12 base odorants and supports iterative refinement through natural-language feedback.","excerpt":"an AI-powered 12-channel wearable olfactory system that maps free-form natural-language descriptions to 12 base odorants","support":"source-stated"},{"text":"In a study of 60 participants across 50 benchmark smells, participants distinguished generated target smells above chance in a three-alternative forced-choice task.","excerpt":"In a between-subjects study (N=60) across 50 real-world benchmark smells, participants distinguished AromaGen-generated target smells above chance in a three-alternative forced-choice task","support":"source-stated"},{"text":"Retrieval-augmented AI composition performed comparably to human-expert composition and zero-shot AI composition.","excerpt":"retrieval-augmented AI composition performing comparably to human-expert composition and zero-shot AI composition","support":"source-stated"},{"text":"Natural-language feedback improved perceived similarity for both AI- and human-expert compositions.","excerpt":"Natural-language feedback significantly improved the perceived similarity of both AI- and human-expert compositions.","support":"source-stated"},{"text":"The authors describe the findings as evidence of feasibility for generating perceptible olfactory experiences from natural language using pretrained multimodal AI and human aroma composition data.","excerpt":"Our findings demonstrate the feasibility of using pretrained multimodal AI, with human aroma composition data, to generate perceptible olfactory experiences from natural language.","support":"source-stated"}],"limitations":["The abstract reports a between-subjects study with 60 participants and 50 real-world benchmark smells; it does not establish how results extend beyond those tested smells or study conditions."],"mediaId":"dd79c9c89b446eb5990a4f0a7e0f847649944c83381a9815b61d4abc3eb65065","correctionReason":null,"id":"c73ca7922113b0d95a73ad616b72c32c267754f9fdb172974f2845a333427c70","publicationId":"932bb49c-488f-46ee-854b-2f8e188e90aa","version":1,"sourceUrl":"https://arxiv.org/abs/2604.01650","sourceTitle":"AromaGen: Interactive Generation of Rich Olfactory Experiences with Multimodal Language Models","publisher":"arXiv cs.AI","publishedAt":null,"originalPublished":{"value":"2026-04-02","precision":"day","basis":"publisher_citation_date"},"firstObservedAt":"2026-11-05T05:09:31.293Z","createdAt":"2026-11-10T00:32:54.762Z","historical":true,"entityIds":["arxiv"],"topicIds":["intelligence-and-discovery"],"media":[{"id":"dd79c9c89b446eb5990a4f0a7e0f847649944c83381a9815b61d4abc3eb65065","kind":"article","url":"https://arxiv.org/abs/2604.01650","sourceUrl":"https://arxiv.org/abs/2604.01650","label":"Read original","provenance":"source_link","verification":{"status":"verified","checkedAt":"2026-11-10T00:31:48.903Z","method":"acquired_source","contentType":"text/html"},"recommendationReason":"Recommended by Jev for understanding this source; checked by Luna."}],"sourceMedia":{"status":"not_found","sourceUrl":"https://arxiv.org/abs/2604.01650","checkedAt":"2026-11-10T00:48:07.671Z","items":[],"note":"No suitable media was found by the supported page extraction."},"editorialDecision":{"outcome":"accept","reasonCodes":["material_development"],"rubricVersion":"stream-admission-v5","mediaId":"dd79c9c89b446eb5990a4f0a7e0f847649944c83381a9815b61d4abc3eb65065","confidence":0.81,"provider":"typesafe","model":"jev-1.13.0","inputHash":"cb2e1335e209abcb34f0943b661f7b25bb30bcb9c39404a3249f7f9dd7abe4b6","decidedAt":"2026-11-10T00:31:50.478Z"},"provenance":{"textHash":"ff54a1ff980e8a6f56a847c6fc5b07747c643ad62225fb0bd95ed90df40a28fa","observationId":"f5be1bb2ff566396e58573e83a8c302d377bbe104a894dbc1d00a7477b74415b","retrievedAt":"2026-11-10T00:31:48.548Z","kind":"publisher-page","preparer":{"model":"gpt-6-luna","runId":"b2f9fdfd-f735-4c03-8823-fa2a039fc837"},"reviewer":{"model":"gpt-6-luna","runId":"25dbb86f-e1e1-4b88-8545-828fa0bf4d6f"}},"audit":{"status":"passed","reviewedAt":"2026-11-10T00:32:54.762Z","model":"anthropic/claude-opus-5.5"},"sourceAttribution":{"authors":[{"name":"Chen, Awu","url":null,"kind":"unknown"},{"name":"Wang, Silang","url":null,"kind":"unknown"},{"name":"Wen, Yunge","url":null,"kind":"unknown"},{"name":"Yu, Jianing","url":null,"kind":"unknown"},{"name":"Eisa, Zinab","url":null,"kind":"unknown"},{"name":"Lopes, Ana Julia Santos","url":null,"kind":"unknown"},{"name":"Brooks, Jas","url":null,"kind":"unknown"},{"name":"Ishii, Hiroshi","url":null,"kind":"unknown"},{"name":"Liang, Paul Pu","url":null,"kind":"unknown"}],"authorRole":"article-author","hosts":[],"status":"stated","authorRoster":{"truncated":false,"omitted":0,"assertionGroupsOmitted":0},"participation":[{"actor":{"name":"Chen, Awu","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Chen, Awu","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2604.01650","locator":"head/meta[name=author|citation_author]"},"assertionId":"7e6f91310d53dc9cf2f9f272e38184c12ff1444ab7a9513506c431628d0f11f7:participation:0","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-10T00:48:07.671Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"7e6f91310d53dc9cf2f9f272e38184c12ff1444ab7a9513506c431628d0f11f7","sourceUrl":"https://arxiv.org/abs/2604.01650","contentSha256":"c79599ac4d84bc092c1625ffd238a6185f26d479ddf52b938f4ae8edfd09ab0f","textHash":"ff54a1ff980e8a6f56a847c6fc5b07747c643ad62225fb0bd95ed90df40a28fa","observedAt":"2026-11-10T00:48:07.671Z"}},{"actor":{"name":"Wang, Silang","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Wang, Silang","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2604.01650","locator":"head/meta[name=author|citation_author]"},"assertionId":"7e6f91310d53dc9cf2f9f272e38184c12ff1444ab7a9513506c431628d0f11f7:participation:1","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-10T00:48:07.671Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"7e6f91310d53dc9cf2f9f272e38184c12ff1444ab7a9513506c431628d0f11f7","sourceUrl":"https://arxiv.org/abs/2604.01650","contentSha256":"c79599ac4d84bc092c1625ffd238a6185f26d479ddf52b938f4ae8edfd09ab0f","textHash":"ff54a1ff980e8a6f56a847c6fc5b07747c643ad62225fb0bd95ed90df40a28fa","observedAt":"2026-11-10T00:48:07.671Z"}},{"actor":{"name":"Wen, Yunge","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Wen, Yunge","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2604.01650","locator":"head/meta[name=author|citation_author]"},"assertionId":"7e6f91310d53dc9cf2f9f272e38184c12ff1444ab7a9513506c431628d0f11f7:participation:2","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-10T00:48:07.671Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"7e6f91310d53dc9cf2f9f272e38184c12ff1444ab7a9513506c431628d0f11f7","sourceUrl":"https://arxiv.org/abs/2604.01650","contentSha256":"c79599ac4d84bc092c1625ffd238a6185f26d479ddf52b938f4ae8edfd09ab0f","textHash":"ff54a1ff980e8a6f56a847c6fc5b07747c643ad62225fb0bd95ed90df40a28fa","observedAt":"2026-11-10T00:48:07.671Z"}},{"actor":{"name":"Yu, Jianing","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Yu, Jianing","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2604.01650","locator":"head/meta[name=author|citation_author]"},"assertionId":"7e6f91310d53dc9cf2f9f272e38184c12ff1444ab7a9513506c431628d0f11f7:participation:3","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-10T00:48:07.671Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"7e6f91310d53dc9cf2f9f272e38184c12ff1444ab7a9513506c431628d0f11f7","sourceUrl":"https://arxiv.org/abs/2604.01650","contentSha256":"c79599ac4d84bc092c1625ffd238a6185f26d479ddf52b938f4ae8edfd09ab0f","textHash":"ff54a1ff980e8a6f56a847c6fc5b07747c643ad62225fb0bd95ed90df40a28fa","observedAt":"2026-11-10T00:48:07.671Z"}},{"actor":{"name":"Eisa, Zinab","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Eisa, Zinab","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2604.01650","locator":"head/meta[name=author|citation_author]"},"assertionId":"7e6f91310d53dc9cf2f9f272e38184c12ff1444ab7a9513506c431628d0f11f7:participation:4","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-10T00:48:07.671Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"7e6f91310d53dc9cf2f9f272e38184c12ff1444ab7a9513506c431628d0f11f7","sourceUrl":"https://arxiv.org/abs/2604.01650","contentSha256":"c79599ac4d84bc092c1625ffd238a6185f26d479ddf52b938f4ae8edfd09ab0f","textHash":"ff54a1ff980e8a6f56a847c6fc5b07747c643ad62225fb0bd95ed90df40a28fa","observedAt":"2026-11-10T00:48:07.671Z"}},{"actor":{"name":"Lopes, Ana Julia Santos","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Lopes, Ana Julia Santos","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2604.01650","locator":"head/meta[name=author|citation_author]"},"assertionId":"7e6f91310d53dc9cf2f9f272e38184c12ff1444ab7a9513506c431628d0f11f7:participation:5","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-10T00:48:07.671Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"7e6f91310d53dc9cf2f9f272e38184c12ff1444ab7a9513506c431628d0f11f7","sourceUrl":"https://arxiv.org/abs/2604.01650","contentSha256":"c79599ac4d84bc092c1625ffd238a6185f26d479ddf52b938f4ae8edfd09ab0f","textHash":"ff54a1ff980e8a6f56a847c6fc5b07747c643ad62225fb0bd95ed90df40a28fa","observedAt":"2026-11-10T00:48:07.671Z"}},{"actor":{"name":"Brooks, Jas","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Brooks, Jas","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2604.01650","locator":"head/meta[name=author|citation_author]"},"assertionId":"7e6f91310d53dc9cf2f9f272e38184c12ff1444ab7a9513506c431628d0f11f7:participation:6","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-10T00:48:07.671Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"7e6f91310d53dc9cf2f9f272e38184c12ff1444ab7a9513506c431628d0f11f7","sourceUrl":"https://arxiv.org/abs/2604.01650","contentSha256":"c79599ac4d84bc092c1625ffd238a6185f26d479ddf52b938f4ae8edfd09ab0f","textHash":"ff54a1ff980e8a6f56a847c6fc5b07747c643ad62225fb0bd95ed90df40a28fa","observedAt":"2026-11-10T00:48:07.671Z"}},{"actor":{"name":"Ishii, Hiroshi","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Ishii, Hiroshi","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2604.01650","locator":"head/meta[name=author|citation_author]"},"assertionId":"7e6f91310d53dc9cf2f9f272e38184c12ff1444ab7a9513506c431628d0f11f7:participation:7","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-10T00:48:07.671Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"7e6f91310d53dc9cf2f9f272e38184c12ff1444ab7a9513506c431628d0f11f7","sourceUrl":"https://arxiv.org/abs/2604.01650","contentSha256":"c79599ac4d84bc092c1625ffd238a6185f26d479ddf52b938f4ae8edfd09ab0f","textHash":"ff54a1ff980e8a6f56a847c6fc5b07747c643ad62225fb0bd95ed90df40a28fa","observedAt":"2026-11-10T00:48:07.671Z"}},{"actor":{"name":"Liang, Paul Pu","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Liang, Paul Pu","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2604.01650","locator":"head/meta[name=author|citation_author]"},"assertionId":"7e6f91310d53dc9cf2f9f272e38184c12ff1444ab7a9513506c431628d0f11f7:participation:8","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-10T00:48:07.671Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"7e6f91310d53dc9cf2f9f272e38184c12ff1444ab7a9513506c431628d0f11f7","sourceUrl":"https://arxiv.org/abs/2604.01650","contentSha256":"c79599ac4d84bc092c1625ffd238a6185f26d479ddf52b938f4ae8edfd09ab0f","textHash":"ff54a1ff980e8a6f56a847c6fc5b07747c643ad62225fb0bd95ed90df40a28fa","observedAt":"2026-11-10T00:48:07.671Z"}}],"participationCoverage":{"total":9,"returned":9,"complete":true,"fullAssertions":"source-metadata-endpoint"},"distribution":null,"publisher":null,"sourceUrl":"https://arxiv.org/abs/2604.01650","snapshotId":"7e6f91310d53dc9cf2f9f272e38184c12ff1444ab7a9513506c431628d0f11f7","contentSha256":"c79599ac4d84bc092c1625ffd238a6185f26d479ddf52b938f4ae8edfd09ab0f","textHash":"ff54a1ff980e8a6f56a847c6fc5b07747c643ad62225fb0bd95ed90df40a28fa","observedAt":"2026-11-10T00:48:07.671Z","basis":"publisher-page-metadata","sourceLinks":53,"linksTruncated":false,"metadataUrl":"/api/pipeline/stream/c73ca7922113b0d95a73ad616b72c32c267754f9fdb172974f2845a333427c70/source"}}}],"audits":[{"publicationId":"932bb49c-488f-46ee-854b-2f8e188e90aa","verdict":"passed","findings":[{"dimension":"attribution","status":"pass","explanation":"The reading attributes the system, study and feasibility claim to the paper's authors on arXiv. The title's 'arXiv cs.AI' label matches the publisher field. The page lists the primary category as cs.HC, with cs.AI as a cross-list, which is a minor framing point and not a misattribution.","excerpt":"Subjects:\nHuman-Computer Interaction (cs.HC); Artificial Intelligence (cs.AI)"},{"dimension":"summary","status":"pass","explanation":"The summary names the actor and the development: a wearable 12-channel system that turns language into odorant mixtures. It explains three-alternative forced-choice and retrieval-augmented composition in plain terms and keeps the authors' feasibility framing. It also says what remains unknown. 'Identified' slightly paraphrases the source's 'distinguished', and 'significantly' is omitted. Neither change alters the meaning. The question follows from the evidence.","excerpt":"participants distinguished AromaGen-generated target smells above chance in a three-alternative forced-choice task"},{"dimension":"date","status":"pass","explanation":"originalPublished is 2026-04-02, which matches the v1 submission date. The later v5 revision (3 Nov 2026) and the feed date (5 Nov 2026) are correctly not used.","excerpt":"[Submitted on 2 Apr 2026 (v1), last revised 3 Nov 2026 (this version, v5)]"},{"dimension":"support","status":"pass","explanation":"Each claim is an exact or close restatement of the abstract and is labeled as source-stated. The limitation about generalizing beyond the 50 benchmark smells is a fair statement of what the abstract does not establish. No benefits or forecasts are invented, and the feasibility claim stays attributed to the authors rather than presented as fact.","excerpt":"Our findings demonstrate the feasibility of using pretrained multimodal AI, with human aroma composition data, to generate perceptible olfactory experiences from natural language."},{"dimension":"media","status":"pass","explanation":"The selected media ID is the verified direct link to the original arXiv abstract page. It is the most direct way for readers to reach the source.","excerpt":null}],"createdAt":"2026-11-10T00:32:54.762Z","model":"anthropic/claude-opus-5.5"}],"correctionUrl":"https://escapemeridian.com/corrections/"}