{"closest_prior_art":[{"name":"Two-stage high-throughput screening with false-discovery-rate control","overlap":"Defines a large screening family, controls false discoveries, advances selected hits to a confirmatory stage, and evaluates the approach using simulations and real screening data under a bounded budget.","remaining_difference":"It concerns chemical discovery, not administrative allegations; it does not preserve an enforcement referral's complete search history, assign action-calibrated evidentiary labels, require an independently assigned confirmation officer, or constrain coercive governmental use.","source_ids":["SRC4"]},{"name":"NIST AI Risk Management Framework Core","overlap":"Calls for risk-scaled governance, documented scope and limitations, traceable testing and metrics, human oversight, independent assessment by personnel outside the front-line development team, privacy review, and deactivation or recovery controls.","remaining_difference":"It governs AI-system risk generally and does not require the full family of attempted compliance screens to determine a selected allegation's statistical status or require fresh evidence before screen-confirmed enforcement status.","source_ids":["SRC3"]},{"name":"CMS Fraud Prevention System investigative workflow","overlap":"Predictive models generate and prioritize alerts for program-integrity analysts, feed existing investigations, and can contribute to consequential administrative actions such as revoking billing privileges or suspending payments.","remaining_difference":"The opened report does not describe familywise or false-discovery control across all searched models and slices, preservation of null looks, exploratory-versus-confirmed allegation labels, or independent fresh-evidence confirmation.","source_ids":["SRC2"]},{"name":"SEC DERA analytics support for enforcement","overlap":"Agency analytics identify outliers, inform decisions to investigate or abandon inquiries, and may provide support for enforcement action while generally remaining one input among others.","remaining_difference":"The report addresses analytics use and impact measurement, not a complete screening-family docket, multiplicity-aware evidentiary status, or a separated confirmation gate before coercive escalation.","source_ids":["SRC1"]}],"contrastive_claim_falsifier":"The claim of distinctness would be falsified by finding an operative administrative-enforcement procedure that already records every material attempted and null screen as one declared family, makes that family determine an allegation's exploratory or multiplicity-adjusted evidentiary status, and withholds screen-confirmed status until a separately assigned reviewer tests a frozen allegation on fresh-period or independently obtained evidence.","contrastive_claim_remaining":"For data-selected administrative allegations that may materially support coercive action, the proposed combination remains contrastive in the opened sources: preserve the complete search family, make family-level error treatment determine evidentiary status, freeze the selected allegation, and require separated fresh evidence before granting screen-confirmed status, while leaving jurisdiction and the ultimate legal judgment with the authorized official.","experiment_id":"eoa_inverse_innovation_exp13_second_slot_policy60_20260806","gates":{"adequate_source_search":{"rationale":"The bounded search covered the proposal directly, data-dredging and multiple-testing terminology, agency analytics products and governance frameworks, and combinations of multiplicity control with staged confirmation. Exactly four opened sources from four publishers were retained, including official government sources and primary research.","source_ids":["SRC1","SRC2","SRC3","SRC4"],"status":"PASS"},"bounded_next_test":{"rationale":"A retrospective review of 20 closed referrals from one six-month program is bounded, read-only, measurable, and reversible. Query logs can be compared with referral records, while two reviewers classify evidentiary status and record agreement, omissions, time, confidentiality failures, and deadline conflicts. Comparable official reviews demonstrate that agency analytics workflows and documentation can be examined, while NIST supports documented and independent assessment.","source_ids":["SRC1","SRC2","SRC3"],"status":"PASS"},"distinct_testable_claim":{"rationale":"The opened sources separately establish enforcement-alert workflows, documented independent AI assessment, and two-stage false-discovery control, but none imposes the claimed administrative-enforcement combination. Its incremental effect is testable through missing-look counts and reviewer agreement on exploratory, adjusted, and independently confirmed labels.","source_ids":["SRC1","SRC2","SRC3","SRC4"],"status":"PASS"},"no_obvious_safety_or_authority_stop":{"rationale":"The first step changes no case outcome and uses existing audit access, with access controls, confidentiality monitoring, mandatory-deadline checks, and a defined halt-and-disable path. NIST's framework supports documented human oversight, privacy-risk examination, independent assessment, and deactivation controls. Actual agency-specific privilege, retention, and access rules still require verification but present no obvious categorical stop to the shadow review.","source_ids":["SRC3"],"status":"PASS"},"supported_problem":{"rationale":"Official reports show that agency analytics select outliers or automatic alerts that feed investigations and potentially consequential administrative actions. Primary research shows that large-scale screening can suffer high testing-error rates and benefits from false-discovery control and staged confirmation. Direct evidence that administrative referrals omit material null looks and are consequently misread as pre-specified was not found, so support is partial rather than complete.","source_ids":["SRC1","SRC2","SRC4"],"status":"PASS"}},"prior_art_disposition":"ADJACENT_PRIOR_ART","problem_evidence":{"finding":"The consequential screening pathway is visible: SEC analytics have identified outliers leading to an enforcement investigation and may support action, while CMS predictive alerts are prioritized within investigative processes and can contribute to revocations, payment reviews, or suspensions. The statistical selection risk is supported in another screening domain. The retained sources do not directly document the proposal's crucial empirical allegation that unselected administrative looks disappear from referrals and inflate supervisors' assessment of a selected anomaly.","source_ids":["SRC1","SRC2","SRC4"],"status":"PARTLY_SUPPORTED"},"research_id":"eoa_inverse_innovation_exp13_light_screen_20260806","schema_version":1,"screen_id":"E13P073","screen_survival":true,"search_lanes":{"component_combination":{"no_result_note":null,"queries":["\"multiple testing\" fraud detection anomaly screening holdout validation","\"false discovery rate\" fraud detection compliance screening","government compliance analytics \"holdout\" confirmation investigation","false discovery rate two-stage screening confirmatory follow-up"],"source_ids":["SRC3","SRC4"]},"direct_problem_and_intervention":{"no_result_note":"Exact searches for a \"screening-family docket\" and an administrative-enforcement false-discovery procedure produced no relevant direct match; the retained agency sources instead document the underlying analytics-to-investigation workflow.","queries":["\"screening family\" docket anomaly enforcement","administrative enforcement referral \"false discovery rate\"","administrative enforcement data mining multiple comparisons false positive government agency","site:sec.gov enforcement data analytics anomalies referrals investigation false positives"],"source_ids":["SRC1","SRC2"]},"products_practices_and_standards":{"no_result_note":null,"queries":["site:gao.gov predictive analytics enforcement fraud false positives agency investigation","government algorithm enforcement anomaly detection audit trail human review official","site:nist.gov AI RMF documentation monitoring independent evaluation government enforcement","regulatory compliance screening analytics independent confirmation enforcement leads"],"source_ids":["SRC1","SRC2","SRC3"]},"synonyms_and_historical_terms":{"no_result_note":null,"queries":["\"data dredging\" enforcement agency analytics","\"multiple comparisons\" \"enforcement\" analytics government","\"false discovery rate\" fraud detection compliance screening","data snooping selective inference anomaly detection false positives"],"source_ids":["SRC4"]}},"sources":[{"claims_supported":["SEC DERA analytics identified outliers that led to at least one Division of Enforcement investigation.","SEC personnel reported that analytics contributed both to abandoning inquiries and to identifying indicia that may support enforcement action.","DERA management stated that analytics are rarely the sole input to SEC decisions."],"publisher":"U.S. Securities and Exchange Commission, Office of Inspector General","source_id":"SRC1","source_type":"OTHER","title":"Although Highly Valued by End Users, DERA Could Improve Its Analytics Support by Formally Measuring Impact, Where Possible, Report No. 553","url":"https://www.sec.gov/files/Although-Highly-Valued-by-End-Users-DERA-Could-Improve-Report-No-553_0.pdf"},{"claims_supported":["CMS's Fraud Prevention System generates automatic alerts on claims and providers for prioritization by program-integrity analysts.","The alerts were incorporated into existing investigative processes and directed toward administrative actions including revocation of billing privileges.","GAO initially found that CMS lacked an adequate approach for measuring the system's effectiveness, illustrating an oversight gap without establishing a multiplicity gap."],"publisher":"U.S. Government Accountability Office","source_id":"SRC2","source_type":"OTHER","title":"Medicare Fraud Prevention: CMS Has Implemented a Predictive Analytics System, but Needs to Define Measures to Determine Its Effectiveness","url":"https://www.gao.gov/products/gao-13-104"},{"claims_supported":["The framework calls for documented scope, risks, system limitations, human oversight, test sets, metrics, and evaluation results.","It recommends involvement of internal experts outside the front-line development team or independent assessors.","It addresses risk-scaled controls, privacy examination, traceability, monitoring, incident response, and deactivation of systems with inconsistent outcomes."],"publisher":"National Institute of Standards and Technology","source_id":"SRC3","source_type":"OFFICIAL_GUIDANCE","title":"AI Risk Management Framework Core","url":"https://airc.nist.gov/airmf-resources/airmf/5-sec-core/"},{"claims_supported":["Large-scale high-throughput screening can suffer high testing-error rates and high confirmation costs.","The authors propose a two-stage procedure that controls false discovery rate in the confirmatory stage while allocating a bounded budget.","Simulation and real chemical-screening data show the procedure can control error while retaining useful detection power."],"publisher":"arXiv, Cornell University","source_id":"SRC4","source_type":"PRIMARY_RESEARCH","title":"Optimal Design for High-Throughput Screening via False Discovery Rate Control","url":"https://arxiv.org/abs/1707.03462"}],"world_novelty_boundary":"This light, bounded public-web search found no exact opened match for the complete administrative-enforcement docket and confirmation rule. It establishes only an adjacent-prior-art disposition for coarse researchability; it cannot establish world novelty, patentability, market size, expert acceptance, implementation feasibility across agencies, or realized public value."}