{"id":"7d67b7a34d4a30e5d08bc94ec00b5f7d628d54a17690f6bc0df866fe0021084c","title":"arXiv cs.AI: DiscoveryAgent: A Framework for Scientist-centric Agentic AI","argument":"The authors present DiscoveryAgent, a domain-agnostic system that combines autonomous research with scientists’ authority to steer hypotheses between research rounds. Its design addresses a central tension: letting AI investigate while keeping its work inspectable and redirectable. Audit agents examine checkpointed records to trace claims, reproduce analyses, and inform scientists’ next decisions.\n\nAcross two predictive modeling tasks, the authors report results against evaluated baselines: DiscoveryAgent produced the best variable-star classifier among them, and its scientist-steered protein variant-effect predictor was surpassed by only one of 106 ProteinGym baselines. The authors also report that scientists working with audit agents identified 14x more valid errors than scientists performing manual review. Three domain-expert case studies span physical science, social science, and life science. In the Florida case study, the authors report an 83% rise in homeowners-insurance premiums from 2013 to 2023, with a growing, regressive burden. The paper argues that this approach can expand scientists’ empirical reach while preserving their authority over research direction and their ability to challenge evidence. The abstract does not establish how well these results generalize beyond the reported tasks and case studies.","topicIds":["intelligence-and-discovery"],"domain":"science","claimType":"attributed reading","limits":"The supplied evidence is the arXiv abstract, not the full paper; it does not provide the methods or detailed evaluation needed to assess the reported comparisons. The abstract reports results from two predictive modeling tasks and three case studies; it does not establish how well those results generalize to other research settings.","speaker":{"name":"Queen, Owen, Ciuca, Ioana, Dann, Emma, Somalwar, Jean, You, Carine, Dillmann, Steven, Manwadkar, Viraj, Gorrell, Evan, Zhang, Harrison G, Pritchard, Jonathan K, Wechsler, Risa H, Zou, James","roleAtSource":"Article author"},"source":{"title":"DiscoveryAgent: A Framework for Scientist-centric Agentic AI","publisher":"arXiv cs.AI","format":"Article","url":"https://arxiv.org/abs/2611.21950","publishedAt":"2026-11-16","dateBasis":"Date basis not provided."},"evidence":[{"text":"The authors describe DiscoveryAgent as combining autonomous investigation with scientist-directed hypothesis steering, with scientists able to select, refine, or replace hypotheses between research rounds.","excerpt":"Scientists select, refine, or replace hypotheses between research rounds","url":"https://arxiv.org/abs/2611.21950","relationship":"Source-stated"},{"text":"The system uses audit agents to examine checkpointed records, trace claims, reproduce analyses, and inform later decisions.","excerpt":"audit agents interrogate checkpointed records to trace claims, reproduce analyses, and inform subsequent decisions.","url":"https://arxiv.org/abs/2611.21950","relationship":"Source-stated"},{"text":"Across two predictive modeling tasks, the authors report that DiscoveryAgent produced the best variable-star classifier among evaluated baselines and a protein variant-effect predictor surpassed by only one of 106 ProteinGym baselines when scientists steered it.","excerpt":"with scientist steering, a protein variant-effect predictor surpassed by only one of 106 ProteinGym baselines.","url":"https://arxiv.org/abs/2611.21950","relationship":"Source-stated"},{"text":"The authors report that scientists collaborating with audit agents identified 14x more valid errors than scientists performing manual review.","excerpt":"scientists collaborating with audit agents identify 14x more valid errors versus scientists performing manual review.","url":"https://arxiv.org/abs/2611.21950","relationship":"Source-stated"},{"text":"The paper describes three domain-expert case studies across physical science, social science, and life science, including a photometric-redshift estimator, an analysis of Florida homeowners-insurance premiums, and identification of CD4+ T-cell relay genes.","excerpt":"In three case studies led by domain experts across physical science, social science, and life science","url":"https://arxiv.org/abs/2611.21950","relationship":"Source-stated"},{"text":"The authors argue that autonomous research can expand scientists’ empirical reach while keeping scientists in control of research direction and able to understand and challenge the evidence.","excerpt":"autonomous research can expand scientists' empirical reach while centering their authority over research direction and their ability to understand and challenge the evidence.","url":"https://arxiv.org/abs/2611.21950","relationship":"Source-stated"}],"sourceMedia":{"status":"not_found","sourceUrl":"https://arxiv.org/abs/2611.21950","checkedAt":"2026-11-18T18:50:10.719Z","items":[],"note":"No suitable media was found by the supported page extraction."},"assumptions":[],"openQuestion":{"attribution":"Open question","text":"How well do DiscoveryAgent’s reported results generalize beyond the two predictive modeling tasks and three case studies?"},"whatWouldChangeView":{"attribution":"Change condition not recorded","text":""},"editorialComparisons":[],"live":{"title":"arXiv cs.AI: DiscoveryAgent: A Framework for Scientist-centric Agentic AI","summary":"The authors present DiscoveryAgent, a domain-agnostic system that combines autonomous research with scientists’ authority to steer hypotheses between research rounds. Its design addresses a central tension: letting AI investigate while keeping its work inspectable and redirectable. Audit agents examine checkpointed records to trace claims, reproduce analyses, and inform scientists’ next decisions.\n\nAcross two predictive modeling tasks, the authors report results against evaluated baselines: DiscoveryAgent produced the best variable-star classifier among them, and its scientist-steered protein variant-effect predictor was surpassed by only one of 106 ProteinGym baselines. The authors also report that scientists working with audit agents identified 14x more valid errors than scientists performing manual review. Three domain-expert case studies span physical science, social science, and life science. In the Florida case study, the authors report an 83% rise in homeowners-insurance premiums from 2013 to 2023, with a growing, regressive burden. The paper argues that this approach can expand scientists’ empirical reach while preserving their authority over research direction and their ability to challenge evidence. The abstract does not establish how well these results generalize beyond the reported tasks and case studies.","question":"How well do DiscoveryAgent’s reported results generalize beyond the two predictive modeling tasks and three case studies?","claims":[{"text":"The authors describe DiscoveryAgent as combining autonomous investigation with scientist-directed hypothesis steering, with scientists able to select, refine, or replace hypotheses between research rounds.","excerpt":"Scientists select, refine, or replace hypotheses between research rounds","support":"source-stated"},{"text":"The system uses audit agents to examine checkpointed records, trace claims, reproduce analyses, and inform later decisions.","excerpt":"audit agents interrogate checkpointed records to trace claims, reproduce analyses, and inform subsequent decisions.","support":"source-stated"},{"text":"Across two predictive modeling tasks, the authors report that DiscoveryAgent produced the best variable-star classifier among evaluated baselines and a protein variant-effect predictor surpassed by only one of 106 ProteinGym baselines when scientists steered it.","excerpt":"with scientist steering, a protein variant-effect predictor surpassed by only one of 106 ProteinGym baselines.","support":"source-stated"},{"text":"The authors report that scientists collaborating with audit agents identified 14x more valid errors than scientists performing manual review.","excerpt":"scientists collaborating with audit agents identify 14x more valid errors versus scientists performing manual review.","support":"source-stated"},{"text":"The paper describes three domain-expert case studies across physical science, social science, and life science, including a photometric-redshift estimator, an analysis of Florida homeowners-insurance premiums, and identification of CD4+ T-cell relay genes.","excerpt":"In three case studies led by domain experts across physical science, social science, and life science","support":"source-stated"},{"text":"The authors argue that autonomous research can expand scientists’ empirical reach while keeping scientists in control of research direction and able to understand and challenge the evidence.","excerpt":"autonomous research can expand scientists' empirical reach while centering their authority over research direction and their ability to understand and challenge the evidence.","support":"source-stated"}],"limitations":["The supplied evidence is the arXiv abstract, not the full paper; it does not provide the methods or detailed evaluation needed to assess the reported comparisons.","The abstract reports results from two predictive modeling tasks and three case studies; it does not establish how well those results generalize to other research settings."],"mediaId":"798a2217c37623f11d22164bff24402b0124e9157157bfeeb3255e9fd3e5bd2c","correctionReason":null,"id":"7d67b7a34d4a30e5d08bc94ec00b5f7d628d54a17690f6bc0df866fe0021084c","publicationId":"a6f484a4-2f1d-4a0e-8ba2-0609c6838e02","version":1,"sourceUrl":"https://arxiv.org/abs/2611.21950","sourceTitle":"DiscoveryAgent: A Framework for Scientist-centric Agentic AI","publisher":"arXiv cs.AI","publishedAt":null,"originalPublished":{"value":"2026-11-16","precision":"day","basis":"publisher_citation_date"},"firstObservedAt":"2026-11-18T05:13:39.286Z","createdAt":"2026-11-18T17:47:34.471Z","historical":false,"entityIds":["arxiv"],"topicIds":["intelligence-and-discovery"],"media":[{"id":"798a2217c37623f11d22164bff24402b0124e9157157bfeeb3255e9fd3e5bd2c","kind":"article","url":"https://arxiv.org/abs/2611.21950","sourceUrl":"https://arxiv.org/abs/2611.21950","label":"Read original","provenance":"source_link","verification":{"status":"verified","checkedAt":"2026-11-18T17:34:04.580Z","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/2611.21950","checkedAt":"2026-11-18T18:50:10.719Z","items":[],"note":"No suitable media was found by the supported page extraction."},"editorialDecision":{"outcome":"accept","reasonCodes":["material_development"],"rubricVersion":"stream-admission-v9","mediaId":"798a2217c37623f11d22164bff24402b0124e9157157bfeeb3255e9fd3e5bd2c","confidence":0.65,"provider":"typesafe","model":"jev-1.13.0","inputHash":"825d66911dca5b20b66ed139bc55bb8e81c969a74c97b7b6c1364548304d7ee7","decidedAt":"2026-11-18T17:34:05.703Z"},"provenance":{"textHash":"0968b527bbde28063dc97f87bb0b42deaa281341f71e201f5e0a9a27b8641e46","observationId":"8a2f15aaa4495049bc07353d1b32655b7682584f303547c6a58683714f600ddd","retrievedAt":"2026-11-18T17:34:04.374Z","kind":"publisher-page","preparer":{"model":"gpt-6-luna","runId":"48c68356-91a8-4d6b-858b-f60bb1484487"},"reviewer":{"model":"gpt-6-luna","runId":"7068922e-e618-4786-8674-9e42eb388dcd"}},"audit":{"status":"passed","reviewedAt":"2026-11-18T17:47:34.471Z","model":"anthropic/claude-opus-5.5"},"sourceAttribution":{"authors":[{"name":"Queen, Owen","url":null,"kind":"unknown"},{"name":"Ciuca, Ioana","url":null,"kind":"unknown"},{"name":"Dann, Emma","url":null,"kind":"unknown"},{"name":"Somalwar, Jean","url":null,"kind":"unknown"},{"name":"You, Carine","url":null,"kind":"unknown"},{"name":"Dillmann, Steven","url":null,"kind":"unknown"},{"name":"Manwadkar, Viraj","url":null,"kind":"unknown"},{"name":"Gorrell, Evan","url":null,"kind":"unknown"},{"name":"Zhang, Harrison G","url":null,"kind":"unknown"},{"name":"Pritchard, Jonathan K","url":null,"kind":"unknown"},{"name":"Wechsler, Risa H","url":null,"kind":"unknown"},{"name":"Zou, James","url":null,"kind":"unknown"}],"authorRole":"article-author","hosts":[],"status":"stated","authorRoster":{"truncated":false,"omitted":0,"assertionGroupsOmitted":0},"participation":[{"actor":{"name":"Queen, Owen","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Queen, Owen","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2611.21950","locator":"head/meta[name=author|citation_author]"},"assertionId":"36612738a664b4592ec620074a91b2231bf3566e5e3cdcc63ee94ccdaeba0194:participation:0","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-18T18:50:10.719Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"36612738a664b4592ec620074a91b2231bf3566e5e3cdcc63ee94ccdaeba0194","sourceUrl":"https://arxiv.org/abs/2611.21950","contentSha256":"cc5541b6caa7fb9cafe99369f4c1480b6a74e0e746d21c51ea2e920b34b67228","textHash":"0968b527bbde28063dc97f87bb0b42deaa281341f71e201f5e0a9a27b8641e46","observedAt":"2026-11-18T18:50:10.719Z"}},{"actor":{"name":"Ciuca, Ioana","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Ciuca, Ioana","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2611.21950","locator":"head/meta[name=author|citation_author]"},"assertionId":"36612738a664b4592ec620074a91b2231bf3566e5e3cdcc63ee94ccdaeba0194:participation:1","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-18T18:50:10.719Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"36612738a664b4592ec620074a91b2231bf3566e5e3cdcc63ee94ccdaeba0194","sourceUrl":"https://arxiv.org/abs/2611.21950","contentSha256":"cc5541b6caa7fb9cafe99369f4c1480b6a74e0e746d21c51ea2e920b34b67228","textHash":"0968b527bbde28063dc97f87bb0b42deaa281341f71e201f5e0a9a27b8641e46","observedAt":"2026-11-18T18:50:10.719Z"}},{"actor":{"name":"Dann, Emma","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Dann, Emma","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2611.21950","locator":"head/meta[name=author|citation_author]"},"assertionId":"36612738a664b4592ec620074a91b2231bf3566e5e3cdcc63ee94ccdaeba0194:participation:2","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-18T18:50:10.719Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"36612738a664b4592ec620074a91b2231bf3566e5e3cdcc63ee94ccdaeba0194","sourceUrl":"https://arxiv.org/abs/2611.21950","contentSha256":"cc5541b6caa7fb9cafe99369f4c1480b6a74e0e746d21c51ea2e920b34b67228","textHash":"0968b527bbde28063dc97f87bb0b42deaa281341f71e201f5e0a9a27b8641e46","observedAt":"2026-11-18T18:50:10.719Z"}},{"actor":{"name":"Somalwar, Jean","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Somalwar, Jean","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2611.21950","locator":"head/meta[name=author|citation_author]"},"assertionId":"36612738a664b4592ec620074a91b2231bf3566e5e3cdcc63ee94ccdaeba0194:participation:3","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-18T18:50:10.719Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"36612738a664b4592ec620074a91b2231bf3566e5e3cdcc63ee94ccdaeba0194","sourceUrl":"https://arxiv.org/abs/2611.21950","contentSha256":"cc5541b6caa7fb9cafe99369f4c1480b6a74e0e746d21c51ea2e920b34b67228","textHash":"0968b527bbde28063dc97f87bb0b42deaa281341f71e201f5e0a9a27b8641e46","observedAt":"2026-11-18T18:50:10.719Z"}},{"actor":{"name":"You, Carine","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"You, Carine","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2611.21950","locator":"head/meta[name=author|citation_author]"},"assertionId":"36612738a664b4592ec620074a91b2231bf3566e5e3cdcc63ee94ccdaeba0194:participation:4","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-18T18:50:10.719Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"36612738a664b4592ec620074a91b2231bf3566e5e3cdcc63ee94ccdaeba0194","sourceUrl":"https://arxiv.org/abs/2611.21950","contentSha256":"cc5541b6caa7fb9cafe99369f4c1480b6a74e0e746d21c51ea2e920b34b67228","textHash":"0968b527bbde28063dc97f87bb0b42deaa281341f71e201f5e0a9a27b8641e46","observedAt":"2026-11-18T18:50:10.719Z"}},{"actor":{"name":"Dillmann, Steven","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Dillmann, Steven","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2611.21950","locator":"head/meta[name=author|citation_author]"},"assertionId":"36612738a664b4592ec620074a91b2231bf3566e5e3cdcc63ee94ccdaeba0194:participation:5","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-18T18:50:10.719Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"36612738a664b4592ec620074a91b2231bf3566e5e3cdcc63ee94ccdaeba0194","sourceUrl":"https://arxiv.org/abs/2611.21950","contentSha256":"cc5541b6caa7fb9cafe99369f4c1480b6a74e0e746d21c51ea2e920b34b67228","textHash":"0968b527bbde28063dc97f87bb0b42deaa281341f71e201f5e0a9a27b8641e46","observedAt":"2026-11-18T18:50:10.719Z"}},{"actor":{"name":"Manwadkar, Viraj","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Manwadkar, Viraj","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2611.21950","locator":"head/meta[name=author|citation_author]"},"assertionId":"36612738a664b4592ec620074a91b2231bf3566e5e3cdcc63ee94ccdaeba0194:participation:6","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-18T18:50:10.719Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"36612738a664b4592ec620074a91b2231bf3566e5e3cdcc63ee94ccdaeba0194","sourceUrl":"https://arxiv.org/abs/2611.21950","contentSha256":"cc5541b6caa7fb9cafe99369f4c1480b6a74e0e746d21c51ea2e920b34b67228","textHash":"0968b527bbde28063dc97f87bb0b42deaa281341f71e201f5e0a9a27b8641e46","observedAt":"2026-11-18T18:50:10.719Z"}},{"actor":{"name":"Gorrell, Evan","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Gorrell, Evan","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2611.21950","locator":"head/meta[name=author|citation_author]"},"assertionId":"36612738a664b4592ec620074a91b2231bf3566e5e3cdcc63ee94ccdaeba0194:participation:7","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-18T18:50:10.719Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"36612738a664b4592ec620074a91b2231bf3566e5e3cdcc63ee94ccdaeba0194","sourceUrl":"https://arxiv.org/abs/2611.21950","contentSha256":"cc5541b6caa7fb9cafe99369f4c1480b6a74e0e746d21c51ea2e920b34b67228","textHash":"0968b527bbde28063dc97f87bb0b42deaa281341f71e201f5e0a9a27b8641e46","observedAt":"2026-11-18T18:50:10.719Z"}},{"actor":{"name":"Zhang, Harrison G","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Zhang, Harrison G","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2611.21950","locator":"head/meta[name=author|citation_author]"},"assertionId":"36612738a664b4592ec620074a91b2231bf3566e5e3cdcc63ee94ccdaeba0194:participation:8","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-18T18:50:10.719Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"36612738a664b4592ec620074a91b2231bf3566e5e3cdcc63ee94ccdaeba0194","sourceUrl":"https://arxiv.org/abs/2611.21950","contentSha256":"cc5541b6caa7fb9cafe99369f4c1480b6a74e0e746d21c51ea2e920b34b67228","textHash":"0968b527bbde28063dc97f87bb0b42deaa281341f71e201f5e0a9a27b8641e46","observedAt":"2026-11-18T18:50:10.719Z"}},{"actor":{"name":"Pritchard, Jonathan K","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Pritchard, Jonathan K","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2611.21950","locator":"head/meta[name=author|citation_author]"},"assertionId":"36612738a664b4592ec620074a91b2231bf3566e5e3cdcc63ee94ccdaeba0194:participation:9","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-18T18:50:10.719Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"36612738a664b4592ec620074a91b2231bf3566e5e3cdcc63ee94ccdaeba0194","sourceUrl":"https://arxiv.org/abs/2611.21950","contentSha256":"cc5541b6caa7fb9cafe99369f4c1480b6a74e0e746d21c51ea2e920b34b67228","textHash":"0968b527bbde28063dc97f87bb0b42deaa281341f71e201f5e0a9a27b8641e46","observedAt":"2026-11-18T18:50:10.719Z"}},{"actor":{"name":"Wechsler, Risa H","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Wechsler, Risa H","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2611.21950","locator":"head/meta[name=author|citation_author]"},"assertionId":"36612738a664b4592ec620074a91b2231bf3566e5e3cdcc63ee94ccdaeba0194:participation:10","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-18T18:50:10.719Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"36612738a664b4592ec620074a91b2231bf3566e5e3cdcc63ee94ccdaeba0194","sourceUrl":"https://arxiv.org/abs/2611.21950","contentSha256":"cc5541b6caa7fb9cafe99369f4c1480b6a74e0e746d21c51ea2e920b34b67228","textHash":"0968b527bbde28063dc97f87bb0b42deaa281341f71e201f5e0a9a27b8641e46","observedAt":"2026-11-18T18:50:10.719Z"}},{"actor":{"name":"Zou, James","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Zou, James","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2611.21950","locator":"head/meta[name=author|citation_author]"},"assertionId":"36612738a664b4592ec620074a91b2231bf3566e5e3cdcc63ee94ccdaeba0194:participation:11","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-18T18:50:10.719Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"36612738a664b4592ec620074a91b2231bf3566e5e3cdcc63ee94ccdaeba0194","sourceUrl":"https://arxiv.org/abs/2611.21950","contentSha256":"cc5541b6caa7fb9cafe99369f4c1480b6a74e0e746d21c51ea2e920b34b67228","textHash":"0968b527bbde28063dc97f87bb0b42deaa281341f71e201f5e0a9a27b8641e46","observedAt":"2026-11-18T18:50:10.719Z"}}],"participationCoverage":{"total":12,"returned":12,"complete":true,"fullAssertions":"source-metadata-endpoint"},"distribution":null,"publisher":null,"sourceUrl":"https://arxiv.org/abs/2611.21950","snapshotId":"36612738a664b4592ec620074a91b2231bf3566e5e3cdcc63ee94ccdaeba0194","contentSha256":"cc5541b6caa7fb9cafe99369f4c1480b6a74e0e746d21c51ea2e920b34b67228","textHash":"0968b527bbde28063dc97f87bb0b42deaa281341f71e201f5e0a9a27b8641e46","observedAt":"2026-11-18T18:50:10.719Z","basis":"publisher-page-metadata","sourceLinks":52,"linksTruncated":false,"metadataUrl":"/api/pipeline/stream/7d67b7a34d4a30e5d08bc94ec00b5f7d628d54a17690f6bc0df866fe0021084c/source"}},"canonicalUrl":"https://escapemeridian.com/readings/arxiv-cs-ai-discoveryagent-a-framework-for-scientist-centric-7d67b7a3","formats":{"html":"/readings/arxiv-cs-ai-discoveryagent-a-framework-for-scientist-centric-7d67b7a3","markdown":"/readings/arxiv-cs-ai-discoveryagent-a-framework-for-scientist-centric-7d67b7a3.md","json":"/readings/arxiv-cs-ai-discoveryagent-a-framework-for-scientist-centric-7d67b7a3.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":[],"conditions":[],"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-agentic","name":"Agentic AI","explanation":"AI systems that plan and carry out a series of steps toward a goal, often by using other software, rather than giving one answer.","referenceUrl":"https://developers.google.com/machine-learning/glossary#agentic"}],"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":"7d67b7a34d4a30e5d08bc94ec00b5f7d628d54a17690f6bc0df866fe0021084c","revision":"a6f484a4-2f1d-4a0e-8ba2-0609c6838e02","revisionKind":"publication-id","title":"arXiv cs.AI: DiscoveryAgent: A Framework for Scientist-centric Agentic AI","url":"https://escapemeridian.com/readings/arxiv-cs-ai-discoveryagent-a-framework-for-scientist-centric-7d67b7a3"},"source":{"type":"source","id":"https://arxiv.org/abs/2611.21950","revision":"0968b527bbde28063dc97f87bb0b42deaa281341f71e201f5e0a9a27b8641e46","revisionKind":"sha256-source-text","title":"DiscoveryAgent: A Framework for Scientist-centric Agentic AI","url":"https://arxiv.org/abs/2611.21950"},"status":"available","items":[],"next":null,"scope":"Explicit citations in public readings to this reading. Shared topics are not citations. Other record types need an explicit reference adapter."},"revisions":[{"id":"7d67b7a34d4a30e5d08bc94ec00b5f7d628d54a17690f6bc0df866fe0021084c","title":"arXiv cs.AI: DiscoveryAgent: A Framework for Scientist-centric Agentic AI","argument":"The authors present DiscoveryAgent, a domain-agnostic system that combines autonomous research with scientists’ authority to steer hypotheses between research rounds. Its design addresses a central tension: letting AI investigate while keeping its work inspectable and redirectable. Audit agents examine checkpointed records to trace claims, reproduce analyses, and inform scientists’ next decisions.\n\nAcross two predictive modeling tasks, the authors report results against evaluated baselines: DiscoveryAgent produced the best variable-star classifier among them, and its scientist-steered protein variant-effect predictor was surpassed by only one of 106 ProteinGym baselines. The authors also report that scientists working with audit agents identified 14x more valid errors than scientists performing manual review. Three domain-expert case studies span physical science, social science, and life science. In the Florida case study, the authors report an 83% rise in homeowners-insurance premiums from 2013 to 2023, with a growing, regressive burden. The paper argues that this approach can expand scientists’ empirical reach while preserving their authority over research direction and their ability to challenge evidence. The abstract does not establish how well these results generalize beyond the reported tasks and case studies.","topicIds":["intelligence-and-discovery"],"domain":"science","claimType":"attributed reading","limits":"The supplied evidence is the arXiv abstract, not the full paper; it does not provide the methods or detailed evaluation needed to assess the reported comparisons. The abstract reports results from two predictive modeling tasks and three case studies; it does not establish how well those results generalize to other research settings.","speaker":{"name":"Queen, Owen, Ciuca, Ioana, Dann, Emma, Somalwar, Jean, You, Carine, Dillmann, Steven, Manwadkar, Viraj, Gorrell, Evan, Zhang, Harrison G, Pritchard, Jonathan K, Wechsler, Risa H, Zou, James","roleAtSource":"Article author"},"source":{"title":"DiscoveryAgent: A Framework for Scientist-centric Agentic AI","publisher":"arXiv cs.AI","format":"Article","url":"https://arxiv.org/abs/2611.21950","publishedAt":"2026-11-16","dateBasis":"Date basis not provided."},"evidence":[{"text":"The authors describe DiscoveryAgent as combining autonomous investigation with scientist-directed hypothesis steering, with scientists able to select, refine, or replace hypotheses between research rounds.","excerpt":"Scientists select, refine, or replace hypotheses between research rounds","url":"https://arxiv.org/abs/2611.21950","relationship":"Source-stated"},{"text":"The system uses audit agents to examine checkpointed records, trace claims, reproduce analyses, and inform later decisions.","excerpt":"audit agents interrogate checkpointed records to trace claims, reproduce analyses, and inform subsequent decisions.","url":"https://arxiv.org/abs/2611.21950","relationship":"Source-stated"},{"text":"Across two predictive modeling tasks, the authors report that DiscoveryAgent produced the best variable-star classifier among evaluated baselines and a protein variant-effect predictor surpassed by only one of 106 ProteinGym baselines when scientists steered it.","excerpt":"with scientist steering, a protein variant-effect predictor surpassed by only one of 106 ProteinGym baselines.","url":"https://arxiv.org/abs/2611.21950","relationship":"Source-stated"},{"text":"The authors report that scientists collaborating with audit agents identified 14x more valid errors than scientists performing manual review.","excerpt":"scientists collaborating with audit agents identify 14x more valid errors versus scientists performing manual review.","url":"https://arxiv.org/abs/2611.21950","relationship":"Source-stated"},{"text":"The paper describes three domain-expert case studies across physical science, social science, and life science, including a photometric-redshift estimator, an analysis of Florida homeowners-insurance premiums, and identification of CD4+ T-cell relay genes.","excerpt":"In three case studies led by domain experts across physical science, social science, and life science","url":"https://arxiv.org/abs/2611.21950","relationship":"Source-stated"},{"text":"The authors argue that autonomous research can expand scientists’ empirical reach while keeping scientists in control of research direction and able to understand and challenge the evidence.","excerpt":"autonomous research can expand scientists' empirical reach while centering their authority over research direction and their ability to understand and challenge the evidence.","url":"https://arxiv.org/abs/2611.21950","relationship":"Source-stated"}],"sourceMedia":{"status":"not_found","sourceUrl":"https://arxiv.org/abs/2611.21950","checkedAt":"2026-11-18T18:50:10.719Z","items":[],"note":"No suitable media was found by the supported page extraction."},"assumptions":[],"openQuestion":{"attribution":"Open question","text":"How well do DiscoveryAgent’s reported results generalize beyond the two predictive modeling tasks and three case studies?"},"whatWouldChangeView":{"attribution":"Change condition not recorded","text":""},"editorialComparisons":[],"live":{"title":"arXiv cs.AI: DiscoveryAgent: A Framework for Scientist-centric Agentic AI","summary":"The authors present DiscoveryAgent, a domain-agnostic system that combines autonomous research with scientists’ authority to steer hypotheses between research rounds. Its design addresses a central tension: letting AI investigate while keeping its work inspectable and redirectable. Audit agents examine checkpointed records to trace claims, reproduce analyses, and inform scientists’ next decisions.\n\nAcross two predictive modeling tasks, the authors report results against evaluated baselines: DiscoveryAgent produced the best variable-star classifier among them, and its scientist-steered protein variant-effect predictor was surpassed by only one of 106 ProteinGym baselines. The authors also report that scientists working with audit agents identified 14x more valid errors than scientists performing manual review. Three domain-expert case studies span physical science, social science, and life science. In the Florida case study, the authors report an 83% rise in homeowners-insurance premiums from 2013 to 2023, with a growing, regressive burden. The paper argues that this approach can expand scientists’ empirical reach while preserving their authority over research direction and their ability to challenge evidence. The abstract does not establish how well these results generalize beyond the reported tasks and case studies.","question":"How well do DiscoveryAgent’s reported results generalize beyond the two predictive modeling tasks and three case studies?","claims":[{"text":"The authors describe DiscoveryAgent as combining autonomous investigation with scientist-directed hypothesis steering, with scientists able to select, refine, or replace hypotheses between research rounds.","excerpt":"Scientists select, refine, or replace hypotheses between research rounds","support":"source-stated"},{"text":"The system uses audit agents to examine checkpointed records, trace claims, reproduce analyses, and inform later decisions.","excerpt":"audit agents interrogate checkpointed records to trace claims, reproduce analyses, and inform subsequent decisions.","support":"source-stated"},{"text":"Across two predictive modeling tasks, the authors report that DiscoveryAgent produced the best variable-star classifier among evaluated baselines and a protein variant-effect predictor surpassed by only one of 106 ProteinGym baselines when scientists steered it.","excerpt":"with scientist steering, a protein variant-effect predictor surpassed by only one of 106 ProteinGym baselines.","support":"source-stated"},{"text":"The authors report that scientists collaborating with audit agents identified 14x more valid errors than scientists performing manual review.","excerpt":"scientists collaborating with audit agents identify 14x more valid errors versus scientists performing manual review.","support":"source-stated"},{"text":"The paper describes three domain-expert case studies across physical science, social science, and life science, including a photometric-redshift estimator, an analysis of Florida homeowners-insurance premiums, and identification of CD4+ T-cell relay genes.","excerpt":"In three case studies led by domain experts across physical science, social science, and life science","support":"source-stated"},{"text":"The authors argue that autonomous research can expand scientists’ empirical reach while keeping scientists in control of research direction and able to understand and challenge the evidence.","excerpt":"autonomous research can expand scientists' empirical reach while centering their authority over research direction and their ability to understand and challenge the evidence.","support":"source-stated"}],"limitations":["The supplied evidence is the arXiv abstract, not the full paper; it does not provide the methods or detailed evaluation needed to assess the reported comparisons.","The abstract reports results from two predictive modeling tasks and three case studies; it does not establish how well those results generalize to other research settings."],"mediaId":"798a2217c37623f11d22164bff24402b0124e9157157bfeeb3255e9fd3e5bd2c","correctionReason":null,"id":"7d67b7a34d4a30e5d08bc94ec00b5f7d628d54a17690f6bc0df866fe0021084c","publicationId":"a6f484a4-2f1d-4a0e-8ba2-0609c6838e02","version":1,"sourceUrl":"https://arxiv.org/abs/2611.21950","sourceTitle":"DiscoveryAgent: A Framework for Scientist-centric Agentic AI","publisher":"arXiv cs.AI","publishedAt":null,"originalPublished":{"value":"2026-11-16","precision":"day","basis":"publisher_citation_date"},"firstObservedAt":"2026-11-18T05:13:39.286Z","createdAt":"2026-11-18T17:47:34.471Z","historical":false,"entityIds":["arxiv"],"topicIds":["intelligence-and-discovery"],"media":[{"id":"798a2217c37623f11d22164bff24402b0124e9157157bfeeb3255e9fd3e5bd2c","kind":"article","url":"https://arxiv.org/abs/2611.21950","sourceUrl":"https://arxiv.org/abs/2611.21950","label":"Read original","provenance":"source_link","verification":{"status":"verified","checkedAt":"2026-11-18T17:34:04.580Z","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/2611.21950","checkedAt":"2026-11-18T18:50:10.719Z","items":[],"note":"No suitable media was found by the supported page extraction."},"editorialDecision":{"outcome":"accept","reasonCodes":["material_development"],"rubricVersion":"stream-admission-v9","mediaId":"798a2217c37623f11d22164bff24402b0124e9157157bfeeb3255e9fd3e5bd2c","confidence":0.65,"provider":"typesafe","model":"jev-1.13.0","inputHash":"825d66911dca5b20b66ed139bc55bb8e81c969a74c97b7b6c1364548304d7ee7","decidedAt":"2026-11-18T17:34:05.703Z"},"provenance":{"textHash":"0968b527bbde28063dc97f87bb0b42deaa281341f71e201f5e0a9a27b8641e46","observationId":"8a2f15aaa4495049bc07353d1b32655b7682584f303547c6a58683714f600ddd","retrievedAt":"2026-11-18T17:34:04.374Z","kind":"publisher-page","preparer":{"model":"gpt-6-luna","runId":"48c68356-91a8-4d6b-858b-f60bb1484487"},"reviewer":{"model":"gpt-6-luna","runId":"7068922e-e618-4786-8674-9e42eb388dcd"}},"audit":{"status":"passed","reviewedAt":"2026-11-18T17:47:34.471Z","model":"anthropic/claude-opus-5.5"},"sourceAttribution":{"authors":[{"name":"Queen, Owen","url":null,"kind":"unknown"},{"name":"Ciuca, Ioana","url":null,"kind":"unknown"},{"name":"Dann, Emma","url":null,"kind":"unknown"},{"name":"Somalwar, Jean","url":null,"kind":"unknown"},{"name":"You, Carine","url":null,"kind":"unknown"},{"name":"Dillmann, Steven","url":null,"kind":"unknown"},{"name":"Manwadkar, Viraj","url":null,"kind":"unknown"},{"name":"Gorrell, Evan","url":null,"kind":"unknown"},{"name":"Zhang, Harrison G","url":null,"kind":"unknown"},{"name":"Pritchard, Jonathan K","url":null,"kind":"unknown"},{"name":"Wechsler, Risa H","url":null,"kind":"unknown"},{"name":"Zou, James","url":null,"kind":"unknown"}],"authorRole":"article-author","hosts":[],"status":"stated","authorRoster":{"truncated":false,"omitted":0,"assertionGroupsOmitted":0},"participation":[{"actor":{"name":"Queen, Owen","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Queen, Owen","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2611.21950","locator":"head/meta[name=author|citation_author]"},"assertionId":"36612738a664b4592ec620074a91b2231bf3566e5e3cdcc63ee94ccdaeba0194:participation:0","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-18T18:50:10.719Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"36612738a664b4592ec620074a91b2231bf3566e5e3cdcc63ee94ccdaeba0194","sourceUrl":"https://arxiv.org/abs/2611.21950","contentSha256":"cc5541b6caa7fb9cafe99369f4c1480b6a74e0e746d21c51ea2e920b34b67228","textHash":"0968b527bbde28063dc97f87bb0b42deaa281341f71e201f5e0a9a27b8641e46","observedAt":"2026-11-18T18:50:10.719Z"}},{"actor":{"name":"Ciuca, Ioana","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Ciuca, Ioana","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2611.21950","locator":"head/meta[name=author|citation_author]"},"assertionId":"36612738a664b4592ec620074a91b2231bf3566e5e3cdcc63ee94ccdaeba0194:participation:1","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-18T18:50:10.719Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"36612738a664b4592ec620074a91b2231bf3566e5e3cdcc63ee94ccdaeba0194","sourceUrl":"https://arxiv.org/abs/2611.21950","contentSha256":"cc5541b6caa7fb9cafe99369f4c1480b6a74e0e746d21c51ea2e920b34b67228","textHash":"0968b527bbde28063dc97f87bb0b42deaa281341f71e201f5e0a9a27b8641e46","observedAt":"2026-11-18T18:50:10.719Z"}},{"actor":{"name":"Dann, Emma","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Dann, Emma","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2611.21950","locator":"head/meta[name=author|citation_author]"},"assertionId":"36612738a664b4592ec620074a91b2231bf3566e5e3cdcc63ee94ccdaeba0194:participation:2","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-18T18:50:10.719Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"36612738a664b4592ec620074a91b2231bf3566e5e3cdcc63ee94ccdaeba0194","sourceUrl":"https://arxiv.org/abs/2611.21950","contentSha256":"cc5541b6caa7fb9cafe99369f4c1480b6a74e0e746d21c51ea2e920b34b67228","textHash":"0968b527bbde28063dc97f87bb0b42deaa281341f71e201f5e0a9a27b8641e46","observedAt":"2026-11-18T18:50:10.719Z"}},{"actor":{"name":"Somalwar, Jean","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Somalwar, Jean","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2611.21950","locator":"head/meta[name=author|citation_author]"},"assertionId":"36612738a664b4592ec620074a91b2231bf3566e5e3cdcc63ee94ccdaeba0194:participation:3","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-18T18:50:10.719Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"36612738a664b4592ec620074a91b2231bf3566e5e3cdcc63ee94ccdaeba0194","sourceUrl":"https://arxiv.org/abs/2611.21950","contentSha256":"cc5541b6caa7fb9cafe99369f4c1480b6a74e0e746d21c51ea2e920b34b67228","textHash":"0968b527bbde28063dc97f87bb0b42deaa281341f71e201f5e0a9a27b8641e46","observedAt":"2026-11-18T18:50:10.719Z"}},{"actor":{"name":"You, Carine","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"You, Carine","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2611.21950","locator":"head/meta[name=author|citation_author]"},"assertionId":"36612738a664b4592ec620074a91b2231bf3566e5e3cdcc63ee94ccdaeba0194:participation:4","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-18T18:50:10.719Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"36612738a664b4592ec620074a91b2231bf3566e5e3cdcc63ee94ccdaeba0194","sourceUrl":"https://arxiv.org/abs/2611.21950","contentSha256":"cc5541b6caa7fb9cafe99369f4c1480b6a74e0e746d21c51ea2e920b34b67228","textHash":"0968b527bbde28063dc97f87bb0b42deaa281341f71e201f5e0a9a27b8641e46","observedAt":"2026-11-18T18:50:10.719Z"}},{"actor":{"name":"Dillmann, Steven","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Dillmann, Steven","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2611.21950","locator":"head/meta[name=author|citation_author]"},"assertionId":"36612738a664b4592ec620074a91b2231bf3566e5e3cdcc63ee94ccdaeba0194:participation:5","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-18T18:50:10.719Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"36612738a664b4592ec620074a91b2231bf3566e5e3cdcc63ee94ccdaeba0194","sourceUrl":"https://arxiv.org/abs/2611.21950","contentSha256":"cc5541b6caa7fb9cafe99369f4c1480b6a74e0e746d21c51ea2e920b34b67228","textHash":"0968b527bbde28063dc97f87bb0b42deaa281341f71e201f5e0a9a27b8641e46","observedAt":"2026-11-18T18:50:10.719Z"}},{"actor":{"name":"Manwadkar, Viraj","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Manwadkar, Viraj","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2611.21950","locator":"head/meta[name=author|citation_author]"},"assertionId":"36612738a664b4592ec620074a91b2231bf3566e5e3cdcc63ee94ccdaeba0194:participation:6","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-18T18:50:10.719Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"36612738a664b4592ec620074a91b2231bf3566e5e3cdcc63ee94ccdaeba0194","sourceUrl":"https://arxiv.org/abs/2611.21950","contentSha256":"cc5541b6caa7fb9cafe99369f4c1480b6a74e0e746d21c51ea2e920b34b67228","textHash":"0968b527bbde28063dc97f87bb0b42deaa281341f71e201f5e0a9a27b8641e46","observedAt":"2026-11-18T18:50:10.719Z"}},{"actor":{"name":"Gorrell, Evan","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Gorrell, Evan","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2611.21950","locator":"head/meta[name=author|citation_author]"},"assertionId":"36612738a664b4592ec620074a91b2231bf3566e5e3cdcc63ee94ccdaeba0194:participation:7","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-18T18:50:10.719Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"36612738a664b4592ec620074a91b2231bf3566e5e3cdcc63ee94ccdaeba0194","sourceUrl":"https://arxiv.org/abs/2611.21950","contentSha256":"cc5541b6caa7fb9cafe99369f4c1480b6a74e0e746d21c51ea2e920b34b67228","textHash":"0968b527bbde28063dc97f87bb0b42deaa281341f71e201f5e0a9a27b8641e46","observedAt":"2026-11-18T18:50:10.719Z"}},{"actor":{"name":"Zhang, Harrison G","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Zhang, Harrison G","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2611.21950","locator":"head/meta[name=author|citation_author]"},"assertionId":"36612738a664b4592ec620074a91b2231bf3566e5e3cdcc63ee94ccdaeba0194:participation:8","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-18T18:50:10.719Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"36612738a664b4592ec620074a91b2231bf3566e5e3cdcc63ee94ccdaeba0194","sourceUrl":"https://arxiv.org/abs/2611.21950","contentSha256":"cc5541b6caa7fb9cafe99369f4c1480b6a74e0e746d21c51ea2e920b34b67228","textHash":"0968b527bbde28063dc97f87bb0b42deaa281341f71e201f5e0a9a27b8641e46","observedAt":"2026-11-18T18:50:10.719Z"}},{"actor":{"name":"Pritchard, Jonathan K","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Pritchard, Jonathan K","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2611.21950","locator":"head/meta[name=author|citation_author]"},"assertionId":"36612738a664b4592ec620074a91b2231bf3566e5e3cdcc63ee94ccdaeba0194:participation:9","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-18T18:50:10.719Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"36612738a664b4592ec620074a91b2231bf3566e5e3cdcc63ee94ccdaeba0194","sourceUrl":"https://arxiv.org/abs/2611.21950","contentSha256":"cc5541b6caa7fb9cafe99369f4c1480b6a74e0e746d21c51ea2e920b34b67228","textHash":"0968b527bbde28063dc97f87bb0b42deaa281341f71e201f5e0a9a27b8641e46","observedAt":"2026-11-18T18:50:10.719Z"}},{"actor":{"name":"Wechsler, Risa H","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Wechsler, Risa H","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2611.21950","locator":"head/meta[name=author|citation_author]"},"assertionId":"36612738a664b4592ec620074a91b2231bf3566e5e3cdcc63ee94ccdaeba0194:participation:10","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-18T18:50:10.719Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"36612738a664b4592ec620074a91b2231bf3566e5e3cdcc63ee94ccdaeba0194","sourceUrl":"https://arxiv.org/abs/2611.21950","contentSha256":"cc5541b6caa7fb9cafe99369f4c1480b6a74e0e746d21c51ea2e920b34b67228","textHash":"0968b527bbde28063dc97f87bb0b42deaa281341f71e201f5e0a9a27b8641e46","observedAt":"2026-11-18T18:50:10.719Z"}},{"actor":{"name":"Zou, James","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Zou, James","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2611.21950","locator":"head/meta[name=author|citation_author]"},"assertionId":"36612738a664b4592ec620074a91b2231bf3566e5e3cdcc63ee94ccdaeba0194:participation:11","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-18T18:50:10.719Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"36612738a664b4592ec620074a91b2231bf3566e5e3cdcc63ee94ccdaeba0194","sourceUrl":"https://arxiv.org/abs/2611.21950","contentSha256":"cc5541b6caa7fb9cafe99369f4c1480b6a74e0e746d21c51ea2e920b34b67228","textHash":"0968b527bbde28063dc97f87bb0b42deaa281341f71e201f5e0a9a27b8641e46","observedAt":"2026-11-18T18:50:10.719Z"}}],"participationCoverage":{"total":12,"returned":12,"complete":true,"fullAssertions":"source-metadata-endpoint"},"distribution":null,"publisher":null,"sourceUrl":"https://arxiv.org/abs/2611.21950","snapshotId":"36612738a664b4592ec620074a91b2231bf3566e5e3cdcc63ee94ccdaeba0194","contentSha256":"cc5541b6caa7fb9cafe99369f4c1480b6a74e0e746d21c51ea2e920b34b67228","textHash":"0968b527bbde28063dc97f87bb0b42deaa281341f71e201f5e0a9a27b8641e46","observedAt":"2026-11-18T18:50:10.719Z","basis":"publisher-page-metadata","sourceLinks":52,"linksTruncated":false,"metadataUrl":"/api/pipeline/stream/7d67b7a34d4a30e5d08bc94ec00b5f7d628d54a17690f6bc0df866fe0021084c/source"}}}],"audits":[{"publicationId":"a6f484a4-2f1d-4a0e-8ba2-0609c6838e02","verdict":"passed","findings":[{"dimension":"attribution","status":"pass","explanation":"The reading attributes the system, results and argument to the paper's authors throughout ('the authors report', 'the paper argues'). This matches the arXiv listing, where Owen Queen and colleagues present DiscoveryAgent. No source claim is presented as established fact or as Escape Meridian's own view.","excerpt":"We present DiscoveryAgent, a domain-agnostic, scientist-centric system"},{"dimension":"summary","status":"pass","explanation":"The reading leads with the development and explains the central tension: autonomous AI investigation versus scientists' ability to inspect and redirect it. It describes audit agents by what they do and keeps the abstract's core results and closing argument. It omits the abstract's framing of audit agents as 'a scalable form of oversight' and their higher autonomous error-detection rate. Terms such as ProteinGym and regressive are left unexplained and could use tooltips. Neither gap blocks understanding. The generalization question is a genuine open issue.","excerpt":"balancing autonomous investigation with scientists' ability to inspect and redirect ongoing work remains a central design challenge"},{"dimension":"date","status":"pass","explanation":"The listed original publication date of 2026-11-16 matches the arXiv v1 submission date. The later feed date of 2026-11-18 is correctly not used.","excerpt":"[Submitted on 16 Nov 2026]"},{"dimension":"support","status":"pass","explanation":"Each claim and summary statement maps onto the abstract: hypothesis steering, checkpointed audit records, the variable-star and ProteinGym results, the 14x error finding against manual review, the three case studies, the Florida 83% premium rise, and the authors' concluding argument. Each result is framed as reported by the authors, and the limitations correctly note that only the abstract was supplied. The reading adds no unsupported benefits, costs or forecasts.","excerpt":"scientists collaborating with audit agents identify 14x more valid errors versus scientists performing manual review"},{"dimension":"media","status":"pass","explanation":"The selected media is the verified direct arXiv abstract page for this paper. It is the original source and the most direct access point.","excerpt":null}],"createdAt":"2026-11-18T17:47:34.471Z","model":"anthropic/claude-opus-5.5"}],"correctionUrl":"https://escapemeridian.com/corrections/"}