{"id":"c997cc1c4492609702a3a795f5ba1412b0120f2a6c1a773080f93a80e617029a","title":"arXiv cs.AI: MMCOMET links commonsense knowledge to images for visual reasoning","argument":"The authors of MMCOMET argue that commonsense knowledge used in multimodal systems can be relevant in meaning yet poorly matched to an image. They introduce a large-scale multimodal commonsense knowledge graph, extending ATOMIC2020 by linking commonsense triples to representative images. The resource contains over 900K multimodal tuples across 19 relation types, and the authors establish an evaluation framework spanning diverse vision-language tasks.\n\nThe authors report that multimodal grounding provides more contextually aligned and reliable knowledge than text-only knowledge graphs and LLM-based knowledge generation, particularly in reasoning-intensive scenarios. But they say external knowledge’s effectiveness depends on the task: naive integration can introduce noise in perception-dominated settings. Their proposed adaptive knowledge router selects knowledge paths or bypasses knowledge in harmful settings. MMCOMET thus makes the role of external knowledge conditional: image-grounded knowledge may help in some tasks, while adding knowledge may get in the way in others.","topicIds":["intelligence-and-discovery"],"domain":"science","claimType":"attributed reading","limits":"The supplied arXiv page provides the abstract but not task-by-task results or measurements of the reported gains. The abstract describes the adaptive knowledge router's purpose but does not specify when it bypasses knowledge or quantify how much noise it reduces.","speaker":{"name":"Kim, Jinwoo, Cabral, Rina Carines, Pratama, Dhita Putri, Wang, Eileen, Arnaout, Hiba, Yang, Shuo, Liu, Danyang, Ji, Seungyeon, Yang, Jie, Poon, Josiah, Pan, Jeff Z., Han, Kyungreem, Han, Soyeon Caren","roleAtSource":"Article author"},"source":{"title":"MMCOMET: A Large-Scale Multimodal Commonsense Knowledge Graph for Visual Reasoning","publisher":"arXiv cs.AI","format":"Article","url":"https://arxiv.org/abs/2603.01055","publishedAt":"2026-03-01","dateBasis":"Date basis not provided."},"evidence":[{"text":"The authors introduce MMCOMET, a multimodal commonsense knowledge graph that extends ATOMIC2020 by linking commonsense triples to representative images.","excerpt":"MMCOMET extends ATOMIC2020 by associating each commonsense triple with representative images through a hybrid alignment pipeline","url":"https://arxiv.org/abs/2603.01055","relationship":"Source-stated"},{"text":"The resource contains more than 900K multimodal tuples across 19 relation types.","excerpt":"resulting in over 900K multimodal tuples spanning 19 relation types.","url":"https://arxiv.org/abs/2603.01055","relationship":"Source-stated"},{"text":"The authors report that multimodal grounding provides more contextually aligned and reliable knowledge than text-only knowledge graphs and LLM-based knowledge generation, particularly in reasoning-intensive scenarios.","excerpt":"multimodal grounding provides more contextually aligned and reliable knowledge compared to both text-only knowledge graphs and LLM-based knowledge generation, particularly in reasoning-intensive scenarios.","url":"https://arxiv.org/abs/2603.01055","relationship":"Source-stated"},{"text":"The authors report that external knowledge can be task-dependent and that naive integration can introduce noise in perception-dominated settings; they propose an adaptive knowledge router to select paths or bypass knowledge in harmful settings.","excerpt":"the effectiveness of external knowledge is task-dependent, and naive knowledge integration can introduce noise in perception-dominated settings. To mitigate this, we propose an adaptive knowledge router","url":"https://arxiv.org/abs/2603.01055","relationship":"Source-stated"}],"sourceMedia":{"status":"not_found","sourceUrl":"https://arxiv.org/abs/2603.01055","checkedAt":"2026-11-11T22:46:10.700Z","items":[],"note":"No suitable media was found by the supported page extraction."},"assumptions":[],"openQuestion":{"attribution":"Open question","text":"What task-specific results show when the adaptive knowledge router selects knowledge paths or bypasses knowledge in harmful settings?"},"whatWouldChangeView":{"attribution":"Change condition not recorded","text":""},"editorialComparisons":[],"live":{"title":"arXiv cs.AI: MMCOMET links commonsense knowledge to images for visual reasoning","summary":"The authors of MMCOMET argue that commonsense knowledge used in multimodal systems can be relevant in meaning yet poorly matched to an image. They introduce a large-scale multimodal commonsense knowledge graph, extending ATOMIC2020 by linking commonsense triples to representative images. The resource contains over 900K multimodal tuples across 19 relation types, and the authors establish an evaluation framework spanning diverse vision-language tasks.\n\nThe authors report that multimodal grounding provides more contextually aligned and reliable knowledge than text-only knowledge graphs and LLM-based knowledge generation, particularly in reasoning-intensive scenarios. But they say external knowledge’s effectiveness depends on the task: naive integration can introduce noise in perception-dominated settings. Their proposed adaptive knowledge router selects knowledge paths or bypasses knowledge in harmful settings. MMCOMET thus makes the role of external knowledge conditional: image-grounded knowledge may help in some tasks, while adding knowledge may get in the way in others.","question":"What task-specific results show when the adaptive knowledge router selects knowledge paths or bypasses knowledge in harmful settings?","claims":[{"text":"The authors introduce MMCOMET, a multimodal commonsense knowledge graph that extends ATOMIC2020 by linking commonsense triples to representative images.","excerpt":"MMCOMET extends ATOMIC2020 by associating each commonsense triple with representative images through a hybrid alignment pipeline","support":"source-stated"},{"text":"The resource contains more than 900K multimodal tuples across 19 relation types.","excerpt":"resulting in over 900K multimodal tuples spanning 19 relation types.","support":"source-stated"},{"text":"The authors report that multimodal grounding provides more contextually aligned and reliable knowledge than text-only knowledge graphs and LLM-based knowledge generation, particularly in reasoning-intensive scenarios.","excerpt":"multimodal grounding provides more contextually aligned and reliable knowledge compared to both text-only knowledge graphs and LLM-based knowledge generation, particularly in reasoning-intensive scenarios.","support":"source-stated"},{"text":"The authors report that external knowledge can be task-dependent and that naive integration can introduce noise in perception-dominated settings; they propose an adaptive knowledge router to select paths or bypass knowledge in harmful settings.","excerpt":"the effectiveness of external knowledge is task-dependent, and naive knowledge integration can introduce noise in perception-dominated settings. To mitigate this, we propose an adaptive knowledge router","support":"source-stated"}],"limitations":["The supplied arXiv page provides the abstract but not task-by-task results or measurements of the reported gains.","The abstract describes the adaptive knowledge router's purpose but does not specify when it bypasses knowledge or quantify how much noise it reduces."],"mediaId":"e00ea3e13686c2f1ef5c745e441e2325130d7e6243eba0273e15277abecafa22","correctionReason":null,"id":"c997cc1c4492609702a3a795f5ba1412b0120f2a6c1a773080f93a80e617029a","publicationId":"85ff0af2-0fec-4c76-819e-5dc6e90b35c1","version":1,"sourceUrl":"https://arxiv.org/abs/2603.01055","sourceTitle":"MMCOMET: A Large-Scale Multimodal Commonsense Knowledge Graph for Visual Reasoning","publisher":"arXiv cs.AI","publishedAt":null,"originalPublished":{"value":"2026-03-01","precision":"day","basis":"publisher_citation_date"},"firstObservedAt":"2026-11-11T05:06:43.800Z","createdAt":"2026-11-11T22:31:25.724Z","historical":true,"entityIds":["arxiv"],"topicIds":["intelligence-and-discovery"],"media":[{"id":"e00ea3e13686c2f1ef5c745e441e2325130d7e6243eba0273e15277abecafa22","kind":"article","url":"https://arxiv.org/abs/2603.01055","sourceUrl":"https://arxiv.org/abs/2603.01055","label":"Read original","provenance":"source_link","verification":{"status":"verified","checkedAt":"2026-11-11T22:19:01.719Z","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/2603.01055","checkedAt":"2026-11-11T22:46:10.700Z","items":[],"note":"No suitable media was found by the supported page extraction."},"editorialDecision":{"outcome":"accept","reasonCodes":["material_development"],"rubricVersion":"stream-admission-v5","mediaId":"e00ea3e13686c2f1ef5c745e441e2325130d7e6243eba0273e15277abecafa22","confidence":0.84,"provider":"typesafe","model":"jev-1.13.0","inputHash":"0c3dd91c793a292641ac6d8e6a06a157b0084d3d750f7e35b4c040c6d5613a07","decidedAt":"2026-11-11T22:19:03.539Z"},"provenance":{"textHash":"7652a9b0af77e85e56e8339ee4e96424ad3b6a5dd87f1c2c6327091055d508a1","observationId":"9292ceaa622ef38bae48f6e9cfb4f832419a11f6f43f239514187f435a1d8dbf","retrievedAt":"2026-11-11T22:19:01.463Z","kind":"publisher-page","preparer":{"model":"gpt-6-luna","runId":"57c9c6e7-acf8-47de-8a06-03c802683014"},"reviewer":{"model":"gpt-6-luna","runId":"5eca50ec-511c-42ab-80a7-6bc62f675783"}},"audit":{"status":"passed","reviewedAt":"2026-11-11T22:31:25.724Z","model":"anthropic/claude-opus-5.5"},"sourceAttribution":{"authors":[{"name":"Kim, Jinwoo","url":null,"kind":"unknown"},{"name":"Cabral, Rina Carines","url":null,"kind":"unknown"},{"name":"Pratama, Dhita Putri","url":null,"kind":"unknown"},{"name":"Wang, Eileen","url":null,"kind":"unknown"},{"name":"Arnaout, Hiba","url":null,"kind":"unknown"},{"name":"Yang, Shuo","url":null,"kind":"unknown"},{"name":"Liu, Danyang","url":null,"kind":"unknown"},{"name":"Ji, Seungyeon","url":null,"kind":"unknown"},{"name":"Yang, Jie","url":null,"kind":"unknown"},{"name":"Poon, Josiah","url":null,"kind":"unknown"},{"name":"Pan, Jeff Z.","url":null,"kind":"unknown"},{"name":"Han, Kyungreem","url":null,"kind":"unknown"},{"name":"Han, Soyeon Caren","url":null,"kind":"unknown"}],"authorRole":"article-author","hosts":[],"status":"stated","authorRoster":{"truncated":false,"omitted":0,"assertionGroupsOmitted":0},"participation":[{"actor":{"name":"Kim, Jinwoo","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Kim, Jinwoo","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2603.01055","locator":"head/meta[name=author|citation_author]"},"assertionId":"fdfcc25c1ff2c3cca8e123f4a4ec9c44a0db047e2b30ff6f86b01082a1b395b0:participation:0","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-11T22:46:10.700Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"fdfcc25c1ff2c3cca8e123f4a4ec9c44a0db047e2b30ff6f86b01082a1b395b0","sourceUrl":"https://arxiv.org/abs/2603.01055","contentSha256":"c2717e1312bc7390dde6dc82d8cb2107d50fdeeeab25fb44844c46370368e6ba","textHash":"7652a9b0af77e85e56e8339ee4e96424ad3b6a5dd87f1c2c6327091055d508a1","observedAt":"2026-11-11T22:46:10.700Z"}},{"actor":{"name":"Cabral, Rina Carines","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Cabral, Rina Carines","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2603.01055","locator":"head/meta[name=author|citation_author]"},"assertionId":"fdfcc25c1ff2c3cca8e123f4a4ec9c44a0db047e2b30ff6f86b01082a1b395b0:participation:1","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-11T22:46:10.700Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"fdfcc25c1ff2c3cca8e123f4a4ec9c44a0db047e2b30ff6f86b01082a1b395b0","sourceUrl":"https://arxiv.org/abs/2603.01055","contentSha256":"c2717e1312bc7390dde6dc82d8cb2107d50fdeeeab25fb44844c46370368e6ba","textHash":"7652a9b0af77e85e56e8339ee4e96424ad3b6a5dd87f1c2c6327091055d508a1","observedAt":"2026-11-11T22:46:10.700Z"}},{"actor":{"name":"Pratama, Dhita Putri","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Pratama, Dhita Putri","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2603.01055","locator":"head/meta[name=author|citation_author]"},"assertionId":"fdfcc25c1ff2c3cca8e123f4a4ec9c44a0db047e2b30ff6f86b01082a1b395b0:participation:2","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-11T22:46:10.700Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"fdfcc25c1ff2c3cca8e123f4a4ec9c44a0db047e2b30ff6f86b01082a1b395b0","sourceUrl":"https://arxiv.org/abs/2603.01055","contentSha256":"c2717e1312bc7390dde6dc82d8cb2107d50fdeeeab25fb44844c46370368e6ba","textHash":"7652a9b0af77e85e56e8339ee4e96424ad3b6a5dd87f1c2c6327091055d508a1","observedAt":"2026-11-11T22:46:10.700Z"}},{"actor":{"name":"Wang, Eileen","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Wang, Eileen","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2603.01055","locator":"head/meta[name=author|citation_author]"},"assertionId":"fdfcc25c1ff2c3cca8e123f4a4ec9c44a0db047e2b30ff6f86b01082a1b395b0:participation:3","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-11T22:46:10.700Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"fdfcc25c1ff2c3cca8e123f4a4ec9c44a0db047e2b30ff6f86b01082a1b395b0","sourceUrl":"https://arxiv.org/abs/2603.01055","contentSha256":"c2717e1312bc7390dde6dc82d8cb2107d50fdeeeab25fb44844c46370368e6ba","textHash":"7652a9b0af77e85e56e8339ee4e96424ad3b6a5dd87f1c2c6327091055d508a1","observedAt":"2026-11-11T22:46:10.700Z"}},{"actor":{"name":"Arnaout, Hiba","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Arnaout, Hiba","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2603.01055","locator":"head/meta[name=author|citation_author]"},"assertionId":"fdfcc25c1ff2c3cca8e123f4a4ec9c44a0db047e2b30ff6f86b01082a1b395b0:participation:4","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-11T22:46:10.700Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"fdfcc25c1ff2c3cca8e123f4a4ec9c44a0db047e2b30ff6f86b01082a1b395b0","sourceUrl":"https://arxiv.org/abs/2603.01055","contentSha256":"c2717e1312bc7390dde6dc82d8cb2107d50fdeeeab25fb44844c46370368e6ba","textHash":"7652a9b0af77e85e56e8339ee4e96424ad3b6a5dd87f1c2c6327091055d508a1","observedAt":"2026-11-11T22:46:10.700Z"}},{"actor":{"name":"Yang, Shuo","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Yang, Shuo","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2603.01055","locator":"head/meta[name=author|citation_author]"},"assertionId":"fdfcc25c1ff2c3cca8e123f4a4ec9c44a0db047e2b30ff6f86b01082a1b395b0:participation:5","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-11T22:46:10.700Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"fdfcc25c1ff2c3cca8e123f4a4ec9c44a0db047e2b30ff6f86b01082a1b395b0","sourceUrl":"https://arxiv.org/abs/2603.01055","contentSha256":"c2717e1312bc7390dde6dc82d8cb2107d50fdeeeab25fb44844c46370368e6ba","textHash":"7652a9b0af77e85e56e8339ee4e96424ad3b6a5dd87f1c2c6327091055d508a1","observedAt":"2026-11-11T22:46:10.700Z"}},{"actor":{"name":"Liu, Danyang","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Liu, Danyang","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2603.01055","locator":"head/meta[name=author|citation_author]"},"assertionId":"fdfcc25c1ff2c3cca8e123f4a4ec9c44a0db047e2b30ff6f86b01082a1b395b0:participation:6","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-11T22:46:10.700Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"fdfcc25c1ff2c3cca8e123f4a4ec9c44a0db047e2b30ff6f86b01082a1b395b0","sourceUrl":"https://arxiv.org/abs/2603.01055","contentSha256":"c2717e1312bc7390dde6dc82d8cb2107d50fdeeeab25fb44844c46370368e6ba","textHash":"7652a9b0af77e85e56e8339ee4e96424ad3b6a5dd87f1c2c6327091055d508a1","observedAt":"2026-11-11T22:46:10.700Z"}},{"actor":{"name":"Ji, Seungyeon","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Ji, Seungyeon","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2603.01055","locator":"head/meta[name=author|citation_author]"},"assertionId":"fdfcc25c1ff2c3cca8e123f4a4ec9c44a0db047e2b30ff6f86b01082a1b395b0:participation:7","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-11T22:46:10.700Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"fdfcc25c1ff2c3cca8e123f4a4ec9c44a0db047e2b30ff6f86b01082a1b395b0","sourceUrl":"https://arxiv.org/abs/2603.01055","contentSha256":"c2717e1312bc7390dde6dc82d8cb2107d50fdeeeab25fb44844c46370368e6ba","textHash":"7652a9b0af77e85e56e8339ee4e96424ad3b6a5dd87f1c2c6327091055d508a1","observedAt":"2026-11-11T22:46:10.700Z"}},{"actor":{"name":"Yang, Jie","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Yang, Jie","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2603.01055","locator":"head/meta[name=author|citation_author]"},"assertionId":"fdfcc25c1ff2c3cca8e123f4a4ec9c44a0db047e2b30ff6f86b01082a1b395b0:participation:8","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-11T22:46:10.700Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"fdfcc25c1ff2c3cca8e123f4a4ec9c44a0db047e2b30ff6f86b01082a1b395b0","sourceUrl":"https://arxiv.org/abs/2603.01055","contentSha256":"c2717e1312bc7390dde6dc82d8cb2107d50fdeeeab25fb44844c46370368e6ba","textHash":"7652a9b0af77e85e56e8339ee4e96424ad3b6a5dd87f1c2c6327091055d508a1","observedAt":"2026-11-11T22:46:10.700Z"}},{"actor":{"name":"Poon, Josiah","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Poon, Josiah","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2603.01055","locator":"head/meta[name=author|citation_author]"},"assertionId":"fdfcc25c1ff2c3cca8e123f4a4ec9c44a0db047e2b30ff6f86b01082a1b395b0:participation:9","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-11T22:46:10.700Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"fdfcc25c1ff2c3cca8e123f4a4ec9c44a0db047e2b30ff6f86b01082a1b395b0","sourceUrl":"https://arxiv.org/abs/2603.01055","contentSha256":"c2717e1312bc7390dde6dc82d8cb2107d50fdeeeab25fb44844c46370368e6ba","textHash":"7652a9b0af77e85e56e8339ee4e96424ad3b6a5dd87f1c2c6327091055d508a1","observedAt":"2026-11-11T22:46:10.700Z"}},{"actor":{"name":"Pan, Jeff Z.","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Pan, Jeff Z.","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2603.01055","locator":"head/meta[name=author|citation_author]"},"assertionId":"fdfcc25c1ff2c3cca8e123f4a4ec9c44a0db047e2b30ff6f86b01082a1b395b0:participation:10","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-11T22:46:10.700Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"fdfcc25c1ff2c3cca8e123f4a4ec9c44a0db047e2b30ff6f86b01082a1b395b0","sourceUrl":"https://arxiv.org/abs/2603.01055","contentSha256":"c2717e1312bc7390dde6dc82d8cb2107d50fdeeeab25fb44844c46370368e6ba","textHash":"7652a9b0af77e85e56e8339ee4e96424ad3b6a5dd87f1c2c6327091055d508a1","observedAt":"2026-11-11T22:46:10.700Z"}},{"actor":{"name":"Han, Kyungreem","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Han, Kyungreem","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2603.01055","locator":"head/meta[name=author|citation_author]"},"assertionId":"fdfcc25c1ff2c3cca8e123f4a4ec9c44a0db047e2b30ff6f86b01082a1b395b0:participation:11","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-11T22:46:10.700Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"fdfcc25c1ff2c3cca8e123f4a4ec9c44a0db047e2b30ff6f86b01082a1b395b0","sourceUrl":"https://arxiv.org/abs/2603.01055","contentSha256":"c2717e1312bc7390dde6dc82d8cb2107d50fdeeeab25fb44844c46370368e6ba","textHash":"7652a9b0af77e85e56e8339ee4e96424ad3b6a5dd87f1c2c6327091055d508a1","observedAt":"2026-11-11T22:46:10.700Z"}},{"actor":{"name":"Han, Soyeon Caren","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Han, Soyeon Caren","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2603.01055","locator":"head/meta[name=author|citation_author]"},"assertionId":"fdfcc25c1ff2c3cca8e123f4a4ec9c44a0db047e2b30ff6f86b01082a1b395b0:participation:12","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-11T22:46:10.700Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"fdfcc25c1ff2c3cca8e123f4a4ec9c44a0db047e2b30ff6f86b01082a1b395b0","sourceUrl":"https://arxiv.org/abs/2603.01055","contentSha256":"c2717e1312bc7390dde6dc82d8cb2107d50fdeeeab25fb44844c46370368e6ba","textHash":"7652a9b0af77e85e56e8339ee4e96424ad3b6a5dd87f1c2c6327091055d508a1","observedAt":"2026-11-11T22:46:10.700Z"}}],"participationCoverage":{"total":13,"returned":13,"complete":true,"fullAssertions":"source-metadata-endpoint"},"distribution":null,"publisher":null,"sourceUrl":"https://arxiv.org/abs/2603.01055","snapshotId":"fdfcc25c1ff2c3cca8e123f4a4ec9c44a0db047e2b30ff6f86b01082a1b395b0","contentSha256":"c2717e1312bc7390dde6dc82d8cb2107d50fdeeeab25fb44844c46370368e6ba","textHash":"7652a9b0af77e85e56e8339ee4e96424ad3b6a5dd87f1c2c6327091055d508a1","observedAt":"2026-11-11T22:46:10.700Z","basis":"publisher-page-metadata","sourceLinks":54,"linksTruncated":false,"metadataUrl":"/api/pipeline/stream/c997cc1c4492609702a3a795f5ba1412b0120f2a6c1a773080f93a80e617029a/source"}},"canonicalUrl":"https://escapemeridian.com/readings/arxiv-cs-ai-mmcomet-links-commonsense-knowledge-to-images-for-c997cc1c","formats":{"html":"/readings/arxiv-cs-ai-mmcomet-links-commonsense-knowledge-to-images-for-c997cc1c","markdown":"/readings/arxiv-cs-ai-mmcomet-links-commonsense-knowledge-to-images-for-c997cc1c.md","json":"/readings/arxiv-cs-ai-mmcomet-links-commonsense-knowledge-to-images-for-c997cc1c.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-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"},{"id":"term-llm","name":"Large language model","explanation":"A language model with many learned numerical settings, used to process and generate language.","referenceUrl":"https://developers.google.com/machine-learning/glossary#large-language-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":"c997cc1c4492609702a3a795f5ba1412b0120f2a6c1a773080f93a80e617029a","revision":"85ff0af2-0fec-4c76-819e-5dc6e90b35c1","revisionKind":"publication-id","title":"arXiv cs.AI: MMCOMET links commonsense knowledge to images for visual reasoning","url":"https://escapemeridian.com/readings/arxiv-cs-ai-mmcomet-links-commonsense-knowledge-to-images-for-c997cc1c"},"source":{"type":"source","id":"https://arxiv.org/abs/2603.01055","revision":"7652a9b0af77e85e56e8339ee4e96424ad3b6a5dd87f1c2c6327091055d508a1","revisionKind":"sha256-source-text","title":"MMCOMET: A Large-Scale Multimodal Commonsense Knowledge Graph for Visual Reasoning","url":"https://arxiv.org/abs/2603.01055"},"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":"c997cc1c4492609702a3a795f5ba1412b0120f2a6c1a773080f93a80e617029a","title":"arXiv cs.AI: MMCOMET links commonsense knowledge to images for visual reasoning","argument":"The authors of MMCOMET argue that commonsense knowledge used in multimodal systems can be relevant in meaning yet poorly matched to an image. They introduce a large-scale multimodal commonsense knowledge graph, extending ATOMIC2020 by linking commonsense triples to representative images. The resource contains over 900K multimodal tuples across 19 relation types, and the authors establish an evaluation framework spanning diverse vision-language tasks.\n\nThe authors report that multimodal grounding provides more contextually aligned and reliable knowledge than text-only knowledge graphs and LLM-based knowledge generation, particularly in reasoning-intensive scenarios. But they say external knowledge’s effectiveness depends on the task: naive integration can introduce noise in perception-dominated settings. Their proposed adaptive knowledge router selects knowledge paths or bypasses knowledge in harmful settings. MMCOMET thus makes the role of external knowledge conditional: image-grounded knowledge may help in some tasks, while adding knowledge may get in the way in others.","topicIds":["intelligence-and-discovery"],"domain":"science","claimType":"attributed reading","limits":"The supplied arXiv page provides the abstract but not task-by-task results or measurements of the reported gains. The abstract describes the adaptive knowledge router's purpose but does not specify when it bypasses knowledge or quantify how much noise it reduces.","speaker":{"name":"Kim, Jinwoo, Cabral, Rina Carines, Pratama, Dhita Putri, Wang, Eileen, Arnaout, Hiba, Yang, Shuo, Liu, Danyang, Ji, Seungyeon, Yang, Jie, Poon, Josiah, Pan, Jeff Z., Han, Kyungreem, Han, Soyeon Caren","roleAtSource":"Article author"},"source":{"title":"MMCOMET: A Large-Scale Multimodal Commonsense Knowledge Graph for Visual Reasoning","publisher":"arXiv cs.AI","format":"Article","url":"https://arxiv.org/abs/2603.01055","publishedAt":"2026-03-01","dateBasis":"Date basis not provided."},"evidence":[{"text":"The authors introduce MMCOMET, a multimodal commonsense knowledge graph that extends ATOMIC2020 by linking commonsense triples to representative images.","excerpt":"MMCOMET extends ATOMIC2020 by associating each commonsense triple with representative images through a hybrid alignment pipeline","url":"https://arxiv.org/abs/2603.01055","relationship":"Source-stated"},{"text":"The resource contains more than 900K multimodal tuples across 19 relation types.","excerpt":"resulting in over 900K multimodal tuples spanning 19 relation types.","url":"https://arxiv.org/abs/2603.01055","relationship":"Source-stated"},{"text":"The authors report that multimodal grounding provides more contextually aligned and reliable knowledge than text-only knowledge graphs and LLM-based knowledge generation, particularly in reasoning-intensive scenarios.","excerpt":"multimodal grounding provides more contextually aligned and reliable knowledge compared to both text-only knowledge graphs and LLM-based knowledge generation, particularly in reasoning-intensive scenarios.","url":"https://arxiv.org/abs/2603.01055","relationship":"Source-stated"},{"text":"The authors report that external knowledge can be task-dependent and that naive integration can introduce noise in perception-dominated settings; they propose an adaptive knowledge router to select paths or bypass knowledge in harmful settings.","excerpt":"the effectiveness of external knowledge is task-dependent, and naive knowledge integration can introduce noise in perception-dominated settings. To mitigate this, we propose an adaptive knowledge router","url":"https://arxiv.org/abs/2603.01055","relationship":"Source-stated"}],"sourceMedia":{"status":"not_found","sourceUrl":"https://arxiv.org/abs/2603.01055","checkedAt":"2026-11-11T22:46:10.700Z","items":[],"note":"No suitable media was found by the supported page extraction."},"assumptions":[],"openQuestion":{"attribution":"Open question","text":"What task-specific results show when the adaptive knowledge router selects knowledge paths or bypasses knowledge in harmful settings?"},"whatWouldChangeView":{"attribution":"Change condition not recorded","text":""},"editorialComparisons":[],"live":{"title":"arXiv cs.AI: MMCOMET links commonsense knowledge to images for visual reasoning","summary":"The authors of MMCOMET argue that commonsense knowledge used in multimodal systems can be relevant in meaning yet poorly matched to an image. They introduce a large-scale multimodal commonsense knowledge graph, extending ATOMIC2020 by linking commonsense triples to representative images. The resource contains over 900K multimodal tuples across 19 relation types, and the authors establish an evaluation framework spanning diverse vision-language tasks.\n\nThe authors report that multimodal grounding provides more contextually aligned and reliable knowledge than text-only knowledge graphs and LLM-based knowledge generation, particularly in reasoning-intensive scenarios. But they say external knowledge’s effectiveness depends on the task: naive integration can introduce noise in perception-dominated settings. Their proposed adaptive knowledge router selects knowledge paths or bypasses knowledge in harmful settings. MMCOMET thus makes the role of external knowledge conditional: image-grounded knowledge may help in some tasks, while adding knowledge may get in the way in others.","question":"What task-specific results show when the adaptive knowledge router selects knowledge paths or bypasses knowledge in harmful settings?","claims":[{"text":"The authors introduce MMCOMET, a multimodal commonsense knowledge graph that extends ATOMIC2020 by linking commonsense triples to representative images.","excerpt":"MMCOMET extends ATOMIC2020 by associating each commonsense triple with representative images through a hybrid alignment pipeline","support":"source-stated"},{"text":"The resource contains more than 900K multimodal tuples across 19 relation types.","excerpt":"resulting in over 900K multimodal tuples spanning 19 relation types.","support":"source-stated"},{"text":"The authors report that multimodal grounding provides more contextually aligned and reliable knowledge than text-only knowledge graphs and LLM-based knowledge generation, particularly in reasoning-intensive scenarios.","excerpt":"multimodal grounding provides more contextually aligned and reliable knowledge compared to both text-only knowledge graphs and LLM-based knowledge generation, particularly in reasoning-intensive scenarios.","support":"source-stated"},{"text":"The authors report that external knowledge can be task-dependent and that naive integration can introduce noise in perception-dominated settings; they propose an adaptive knowledge router to select paths or bypass knowledge in harmful settings.","excerpt":"the effectiveness of external knowledge is task-dependent, and naive knowledge integration can introduce noise in perception-dominated settings. To mitigate this, we propose an adaptive knowledge router","support":"source-stated"}],"limitations":["The supplied arXiv page provides the abstract but not task-by-task results or measurements of the reported gains.","The abstract describes the adaptive knowledge router's purpose but does not specify when it bypasses knowledge or quantify how much noise it reduces."],"mediaId":"e00ea3e13686c2f1ef5c745e441e2325130d7e6243eba0273e15277abecafa22","correctionReason":null,"id":"c997cc1c4492609702a3a795f5ba1412b0120f2a6c1a773080f93a80e617029a","publicationId":"85ff0af2-0fec-4c76-819e-5dc6e90b35c1","version":1,"sourceUrl":"https://arxiv.org/abs/2603.01055","sourceTitle":"MMCOMET: A Large-Scale Multimodal Commonsense Knowledge Graph for Visual Reasoning","publisher":"arXiv cs.AI","publishedAt":null,"originalPublished":{"value":"2026-03-01","precision":"day","basis":"publisher_citation_date"},"firstObservedAt":"2026-11-11T05:06:43.800Z","createdAt":"2026-11-11T22:31:25.724Z","historical":true,"entityIds":["arxiv"],"topicIds":["intelligence-and-discovery"],"media":[{"id":"e00ea3e13686c2f1ef5c745e441e2325130d7e6243eba0273e15277abecafa22","kind":"article","url":"https://arxiv.org/abs/2603.01055","sourceUrl":"https://arxiv.org/abs/2603.01055","label":"Read original","provenance":"source_link","verification":{"status":"verified","checkedAt":"2026-11-11T22:19:01.719Z","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/2603.01055","checkedAt":"2026-11-11T22:46:10.700Z","items":[],"note":"No suitable media was found by the supported page extraction."},"editorialDecision":{"outcome":"accept","reasonCodes":["material_development"],"rubricVersion":"stream-admission-v5","mediaId":"e00ea3e13686c2f1ef5c745e441e2325130d7e6243eba0273e15277abecafa22","confidence":0.84,"provider":"typesafe","model":"jev-1.13.0","inputHash":"0c3dd91c793a292641ac6d8e6a06a157b0084d3d750f7e35b4c040c6d5613a07","decidedAt":"2026-11-11T22:19:03.539Z"},"provenance":{"textHash":"7652a9b0af77e85e56e8339ee4e96424ad3b6a5dd87f1c2c6327091055d508a1","observationId":"9292ceaa622ef38bae48f6e9cfb4f832419a11f6f43f239514187f435a1d8dbf","retrievedAt":"2026-11-11T22:19:01.463Z","kind":"publisher-page","preparer":{"model":"gpt-6-luna","runId":"57c9c6e7-acf8-47de-8a06-03c802683014"},"reviewer":{"model":"gpt-6-luna","runId":"5eca50ec-511c-42ab-80a7-6bc62f675783"}},"audit":{"status":"passed","reviewedAt":"2026-11-11T22:31:25.724Z","model":"anthropic/claude-opus-5.5"},"sourceAttribution":{"authors":[{"name":"Kim, Jinwoo","url":null,"kind":"unknown"},{"name":"Cabral, Rina Carines","url":null,"kind":"unknown"},{"name":"Pratama, Dhita Putri","url":null,"kind":"unknown"},{"name":"Wang, Eileen","url":null,"kind":"unknown"},{"name":"Arnaout, Hiba","url":null,"kind":"unknown"},{"name":"Yang, Shuo","url":null,"kind":"unknown"},{"name":"Liu, Danyang","url":null,"kind":"unknown"},{"name":"Ji, Seungyeon","url":null,"kind":"unknown"},{"name":"Yang, Jie","url":null,"kind":"unknown"},{"name":"Poon, Josiah","url":null,"kind":"unknown"},{"name":"Pan, Jeff Z.","url":null,"kind":"unknown"},{"name":"Han, Kyungreem","url":null,"kind":"unknown"},{"name":"Han, Soyeon Caren","url":null,"kind":"unknown"}],"authorRole":"article-author","hosts":[],"status":"stated","authorRoster":{"truncated":false,"omitted":0,"assertionGroupsOmitted":0},"participation":[{"actor":{"name":"Kim, Jinwoo","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Kim, Jinwoo","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2603.01055","locator":"head/meta[name=author|citation_author]"},"assertionId":"fdfcc25c1ff2c3cca8e123f4a4ec9c44a0db047e2b30ff6f86b01082a1b395b0:participation:0","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-11T22:46:10.700Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"fdfcc25c1ff2c3cca8e123f4a4ec9c44a0db047e2b30ff6f86b01082a1b395b0","sourceUrl":"https://arxiv.org/abs/2603.01055","contentSha256":"c2717e1312bc7390dde6dc82d8cb2107d50fdeeeab25fb44844c46370368e6ba","textHash":"7652a9b0af77e85e56e8339ee4e96424ad3b6a5dd87f1c2c6327091055d508a1","observedAt":"2026-11-11T22:46:10.700Z"}},{"actor":{"name":"Cabral, Rina Carines","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Cabral, Rina Carines","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2603.01055","locator":"head/meta[name=author|citation_author]"},"assertionId":"fdfcc25c1ff2c3cca8e123f4a4ec9c44a0db047e2b30ff6f86b01082a1b395b0:participation:1","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-11T22:46:10.700Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"fdfcc25c1ff2c3cca8e123f4a4ec9c44a0db047e2b30ff6f86b01082a1b395b0","sourceUrl":"https://arxiv.org/abs/2603.01055","contentSha256":"c2717e1312bc7390dde6dc82d8cb2107d50fdeeeab25fb44844c46370368e6ba","textHash":"7652a9b0af77e85e56e8339ee4e96424ad3b6a5dd87f1c2c6327091055d508a1","observedAt":"2026-11-11T22:46:10.700Z"}},{"actor":{"name":"Pratama, Dhita Putri","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Pratama, Dhita Putri","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2603.01055","locator":"head/meta[name=author|citation_author]"},"assertionId":"fdfcc25c1ff2c3cca8e123f4a4ec9c44a0db047e2b30ff6f86b01082a1b395b0:participation:2","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-11T22:46:10.700Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"fdfcc25c1ff2c3cca8e123f4a4ec9c44a0db047e2b30ff6f86b01082a1b395b0","sourceUrl":"https://arxiv.org/abs/2603.01055","contentSha256":"c2717e1312bc7390dde6dc82d8cb2107d50fdeeeab25fb44844c46370368e6ba","textHash":"7652a9b0af77e85e56e8339ee4e96424ad3b6a5dd87f1c2c6327091055d508a1","observedAt":"2026-11-11T22:46:10.700Z"}},{"actor":{"name":"Wang, Eileen","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Wang, Eileen","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2603.01055","locator":"head/meta[name=author|citation_author]"},"assertionId":"fdfcc25c1ff2c3cca8e123f4a4ec9c44a0db047e2b30ff6f86b01082a1b395b0:participation:3","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-11T22:46:10.700Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"fdfcc25c1ff2c3cca8e123f4a4ec9c44a0db047e2b30ff6f86b01082a1b395b0","sourceUrl":"https://arxiv.org/abs/2603.01055","contentSha256":"c2717e1312bc7390dde6dc82d8cb2107d50fdeeeab25fb44844c46370368e6ba","textHash":"7652a9b0af77e85e56e8339ee4e96424ad3b6a5dd87f1c2c6327091055d508a1","observedAt":"2026-11-11T22:46:10.700Z"}},{"actor":{"name":"Arnaout, Hiba","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Arnaout, Hiba","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2603.01055","locator":"head/meta[name=author|citation_author]"},"assertionId":"fdfcc25c1ff2c3cca8e123f4a4ec9c44a0db047e2b30ff6f86b01082a1b395b0:participation:4","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-11T22:46:10.700Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"fdfcc25c1ff2c3cca8e123f4a4ec9c44a0db047e2b30ff6f86b01082a1b395b0","sourceUrl":"https://arxiv.org/abs/2603.01055","contentSha256":"c2717e1312bc7390dde6dc82d8cb2107d50fdeeeab25fb44844c46370368e6ba","textHash":"7652a9b0af77e85e56e8339ee4e96424ad3b6a5dd87f1c2c6327091055d508a1","observedAt":"2026-11-11T22:46:10.700Z"}},{"actor":{"name":"Yang, Shuo","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Yang, Shuo","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2603.01055","locator":"head/meta[name=author|citation_author]"},"assertionId":"fdfcc25c1ff2c3cca8e123f4a4ec9c44a0db047e2b30ff6f86b01082a1b395b0:participation:5","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-11T22:46:10.700Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"fdfcc25c1ff2c3cca8e123f4a4ec9c44a0db047e2b30ff6f86b01082a1b395b0","sourceUrl":"https://arxiv.org/abs/2603.01055","contentSha256":"c2717e1312bc7390dde6dc82d8cb2107d50fdeeeab25fb44844c46370368e6ba","textHash":"7652a9b0af77e85e56e8339ee4e96424ad3b6a5dd87f1c2c6327091055d508a1","observedAt":"2026-11-11T22:46:10.700Z"}},{"actor":{"name":"Liu, Danyang","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Liu, Danyang","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2603.01055","locator":"head/meta[name=author|citation_author]"},"assertionId":"fdfcc25c1ff2c3cca8e123f4a4ec9c44a0db047e2b30ff6f86b01082a1b395b0:participation:6","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-11T22:46:10.700Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"fdfcc25c1ff2c3cca8e123f4a4ec9c44a0db047e2b30ff6f86b01082a1b395b0","sourceUrl":"https://arxiv.org/abs/2603.01055","contentSha256":"c2717e1312bc7390dde6dc82d8cb2107d50fdeeeab25fb44844c46370368e6ba","textHash":"7652a9b0af77e85e56e8339ee4e96424ad3b6a5dd87f1c2c6327091055d508a1","observedAt":"2026-11-11T22:46:10.700Z"}},{"actor":{"name":"Ji, Seungyeon","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Ji, Seungyeon","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2603.01055","locator":"head/meta[name=author|citation_author]"},"assertionId":"fdfcc25c1ff2c3cca8e123f4a4ec9c44a0db047e2b30ff6f86b01082a1b395b0:participation:7","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-11T22:46:10.700Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"fdfcc25c1ff2c3cca8e123f4a4ec9c44a0db047e2b30ff6f86b01082a1b395b0","sourceUrl":"https://arxiv.org/abs/2603.01055","contentSha256":"c2717e1312bc7390dde6dc82d8cb2107d50fdeeeab25fb44844c46370368e6ba","textHash":"7652a9b0af77e85e56e8339ee4e96424ad3b6a5dd87f1c2c6327091055d508a1","observedAt":"2026-11-11T22:46:10.700Z"}},{"actor":{"name":"Yang, Jie","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Yang, Jie","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2603.01055","locator":"head/meta[name=author|citation_author]"},"assertionId":"fdfcc25c1ff2c3cca8e123f4a4ec9c44a0db047e2b30ff6f86b01082a1b395b0:participation:8","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-11T22:46:10.700Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"fdfcc25c1ff2c3cca8e123f4a4ec9c44a0db047e2b30ff6f86b01082a1b395b0","sourceUrl":"https://arxiv.org/abs/2603.01055","contentSha256":"c2717e1312bc7390dde6dc82d8cb2107d50fdeeeab25fb44844c46370368e6ba","textHash":"7652a9b0af77e85e56e8339ee4e96424ad3b6a5dd87f1c2c6327091055d508a1","observedAt":"2026-11-11T22:46:10.700Z"}},{"actor":{"name":"Poon, Josiah","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Poon, Josiah","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2603.01055","locator":"head/meta[name=author|citation_author]"},"assertionId":"fdfcc25c1ff2c3cca8e123f4a4ec9c44a0db047e2b30ff6f86b01082a1b395b0:participation:9","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-11T22:46:10.700Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"fdfcc25c1ff2c3cca8e123f4a4ec9c44a0db047e2b30ff6f86b01082a1b395b0","sourceUrl":"https://arxiv.org/abs/2603.01055","contentSha256":"c2717e1312bc7390dde6dc82d8cb2107d50fdeeeab25fb44844c46370368e6ba","textHash":"7652a9b0af77e85e56e8339ee4e96424ad3b6a5dd87f1c2c6327091055d508a1","observedAt":"2026-11-11T22:46:10.700Z"}},{"actor":{"name":"Pan, Jeff Z.","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Pan, Jeff Z.","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2603.01055","locator":"head/meta[name=author|citation_author]"},"assertionId":"fdfcc25c1ff2c3cca8e123f4a4ec9c44a0db047e2b30ff6f86b01082a1b395b0:participation:10","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-11T22:46:10.700Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"fdfcc25c1ff2c3cca8e123f4a4ec9c44a0db047e2b30ff6f86b01082a1b395b0","sourceUrl":"https://arxiv.org/abs/2603.01055","contentSha256":"c2717e1312bc7390dde6dc82d8cb2107d50fdeeeab25fb44844c46370368e6ba","textHash":"7652a9b0af77e85e56e8339ee4e96424ad3b6a5dd87f1c2c6327091055d508a1","observedAt":"2026-11-11T22:46:10.700Z"}},{"actor":{"name":"Han, Kyungreem","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Han, Kyungreem","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2603.01055","locator":"head/meta[name=author|citation_author]"},"assertionId":"fdfcc25c1ff2c3cca8e123f4a4ec9c44a0db047e2b30ff6f86b01082a1b395b0:participation:11","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-11T22:46:10.700Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"fdfcc25c1ff2c3cca8e123f4a4ec9c44a0db047e2b30ff6f86b01082a1b395b0","sourceUrl":"https://arxiv.org/abs/2603.01055","contentSha256":"c2717e1312bc7390dde6dc82d8cb2107d50fdeeeab25fb44844c46370368e6ba","textHash":"7652a9b0af77e85e56e8339ee4e96424ad3b6a5dd87f1c2c6327091055d508a1","observedAt":"2026-11-11T22:46:10.700Z"}},{"actor":{"name":"Han, Soyeon Caren","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Han, Soyeon Caren","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2603.01055","locator":"head/meta[name=author|citation_author]"},"assertionId":"fdfcc25c1ff2c3cca8e123f4a4ec9c44a0db047e2b30ff6f86b01082a1b395b0:participation:12","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-11T22:46:10.700Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"fdfcc25c1ff2c3cca8e123f4a4ec9c44a0db047e2b30ff6f86b01082a1b395b0","sourceUrl":"https://arxiv.org/abs/2603.01055","contentSha256":"c2717e1312bc7390dde6dc82d8cb2107d50fdeeeab25fb44844c46370368e6ba","textHash":"7652a9b0af77e85e56e8339ee4e96424ad3b6a5dd87f1c2c6327091055d508a1","observedAt":"2026-11-11T22:46:10.700Z"}}],"participationCoverage":{"total":13,"returned":13,"complete":true,"fullAssertions":"source-metadata-endpoint"},"distribution":null,"publisher":null,"sourceUrl":"https://arxiv.org/abs/2603.01055","snapshotId":"fdfcc25c1ff2c3cca8e123f4a4ec9c44a0db047e2b30ff6f86b01082a1b395b0","contentSha256":"c2717e1312bc7390dde6dc82d8cb2107d50fdeeeab25fb44844c46370368e6ba","textHash":"7652a9b0af77e85e56e8339ee4e96424ad3b6a5dd87f1c2c6327091055d508a1","observedAt":"2026-11-11T22:46:10.700Z","basis":"publisher-page-metadata","sourceLinks":54,"linksTruncated":false,"metadataUrl":"/api/pipeline/stream/c997cc1c4492609702a3a795f5ba1412b0120f2a6c1a773080f93a80e617029a/source"}}}],"audits":[{"publicationId":"85ff0af2-0fec-4c76-819e-5dc6e90b35c1","verdict":"passed","findings":[{"dimension":"attribution","status":"pass","explanation":"The reading attributes MMCOMET and its reported findings to the paper's authors on arXiv cs.AI. Its comparative claims stay framed as what the authors report or say, not as established fact.","excerpt":"We introduce MMCOMET, a large-scale multimodal commonsense knowledge graph"},{"dimension":"summary","status":"pass","explanation":"The reading opens with the authors' problem: text-only knowledge can match an image in meaning but not visually. It then gives the resource and keeps both the reported benefit and the caveat that naive integration can add noise. Technical terms are kept exactly and can be explained by tooltips. Suggested term-library additions: commonsense triple and ATOMIC2020. The question names a genuine open issue rather than leading the reader. The closing interpretation is hedged and follows from the abstract.","excerpt":"existing resources are predominantly text-only and lack grounding in visual context"},{"dimension":"date","status":"pass","explanation":"originalPublished is 2026-03-01, which matches the v1 submission date. The later revisions (v2 on 2 Nov, v3 on 10 Nov) and the feed date are correctly not used as the original publication date.","excerpt":"[Submitted on 1 Mar 2026 (v1), last revised 10 Nov 2026 (this version, v3)]"},{"dimension":"support","status":"pass","explanation":"Every claim matches the abstract: extending ATOMIC2020, over 900K tuples across 19 relation types, the comparison against text-only knowledge graphs and LLM-based knowledge generation, task dependence, and the router. The closing inference uses may and does not overstate results. The limitations correctly say the abstract gives no quantitative results. The reading does not invent benefits or costs.","excerpt":"we propose an adaptive knowledge router that dynamically selects optimal paths or bypass knowledge in harmful settings"},{"dimension":"media","status":"pass","explanation":"The selected media is the verified direct arXiv abstract page for this paper, which is the most direct way to reach the source.","excerpt":null}],"createdAt":"2026-11-11T22:31:25.724Z","model":"anthropic/claude-opus-5.5"}],"correctionUrl":"https://escapemeridian.com/corrections/"}