{"schema_version":1,"experiment_id":"eoa_inverse_innovation_exp04_retrieval_first_paired20_20260802","cell_id":"invariant_mode_decomposition_design__political_science","round_index":0,"assessments":[{"hypothesis_id":"H1","search_queries":["legislative voting eigenvalue spectral analysis coalition defection network","roll call votes dynamic latent space coalition forecasting legislators","legislative whip vote prediction individual legislators outreach targeting","collective defection mode legislature transition matrix eigenvector"],"sources":[{"source_id":"H1-S1","title":"Voting Behavior, Coalitions and Government Strength through a Complex Network Analysis","publisher":"PLOS ONE / PubMed Central","url":"https://pmc.ncbi.nlm.nih.gov/articles/PMC4280168/","source_class":"PRIMARY_RESEARCH","claims_supported":["Roll-call agreement networks have been used to recover coalition structure and measure government-coalition stability over time.","The method estimates each member's contribution to coalition stability and the impact of possible defection.","The article explicitly contrasts its method with prior spectral-decomposition approaches."]},{"source_id":"H1-S2","title":"Roll Call Vote Prediction with Knowledge Augmented Models","publisher":"Association for Computational Linguistics","url":"https://aclanthology.org/K19-1053/","source_class":"PRIMARY_RESEARCH","claims_supported":["Existing work predicts legislators' roll-call choices from voting histories and other information.","The cited model evaluates predictions at the politician-vote level rather than selecting interventions on collective modes."]},{"source_id":"H1-S3","title":"U.S. House of Representatives Roll Call Votes, 119th Congress, 2nd Session","publisher":"Office of the Clerk, U.S. House of Representatives","url":"https://clerk.house.gov/evs/2026/index.asp","source_class":"OFFICIAL_ORGANIZATION_DATA","claims_supported":["Official roll-call records identify each vote, question, result, and underlying measure.","These records provide observable vote histories but not private grievance signals, concessions, outreach, or whip interventions."]}],"closest_analogue":"Dal Maso et al.'s roll-call network method for tracking government-coalition stability and the impact of individual defections.","overlap":"Both infer collective coalition structure from roll-call behavior, monitor cohesion or stability over time, and seek to identify politically consequential defections rather than relying only on formal party labels.","remaining_difference":"The hypothesis estimates a vote-transition operator, identifies a weakly damped collective grievance combination, and selects concessions or outreach by sensitivity to passage probability. The analogue instead uses agreement-network communities and member contributions; the searched sources do not demonstrate prospective intervention allocation over grievance combinations. This difference is testable by comparing held-out passage and defection performance against the network metric and legislator-level predictors.","classification":"POSSIBLE_DISTINCTION","disposition":"ADVANCE","rationale":"The closest work substantially overlaps on coalition monitoring and defection relevance, but the prospective modal intervention rule is a concrete, falsifiable difference rather than a renamed coalition score."},{"hypothesis_id":"H2","search_queries":["public consultation summary minority views thematic analysis guidance government","public comment analysis topic modeling rare themes reconstruction residual","\"Participatory provenance\" representational auditing public consultation","public consultation summarization representativeness audit omitted dissent"],"sources":[{"source_id":"H2-S1","title":"Participatory provenance as representational auditing for AI-mediated public consultation","publisher":"arXiv","url":"https://arxiv.org/abs/2604.20711","source_class":"PRIMARY_RESEARCH","claims_supported":["The paper introduces a formal framework for measuring how individual consultation submissions are transformed, filtered, or lost in summaries.","It evaluates official Canadian consultation summaries against coverage baselines.","It reports concentrated exclusion of dissenting, skeptical, brief, and semantically isolated contributions and provides an interactive tool for iterative summary auditing."]},{"source_id":"H2-S2","title":"Can AI Truly Represent Your Voice in Deliberations? A Comprehensive Study of Large-Scale Opinion Aggregation with LLMs","publisher":"OpenReview","url":"https://openreview.net/forum?id=bl9hFm04Lc","source_class":"PRIMARY_RESEARCH","claims_supported":["DeliberationBank supplies participant opinions and human judgments of summary representativeness, informativeness, neutrality, and policy approval.","Its evaluation reports systematic underrepresentation of minority viewpoints in deliberation summaries."]},{"source_id":"H2-S3","title":"Consultation principles: guidance","publisher":"UK Cabinet Office / GOV.UK","url":"https://www.gov.uk/government/publications/consultation-principles-guidance/consultation-principles-2018","source_class":"OFFICIAL_GUIDANCE","claims_supported":["Government responses should explain the consultation responses received and how they informed policy.","Consultation reporting is expected to facilitate scrutiny, although the guidance does not prescribe reconstruction residuals."]},{"source_id":"H2-S4","title":"Equality and Human Rights Mainstreaming Strategy: consultation analysis","publisher":"Scottish Government","url":"https://www.gov.scot/binaries/content/documents/govscot/publications/consultation-analysis/2025/07/equality-human-rights-mainstreaming-strategy-consultation-analysis/documents/consultation-equality-human-rights-mainstreaming-strategy-analysis-responses/consultation-equality-human-rights-mainstreaming-strategy-analysis-responses/govscot%253Adocument/consultation-equality-human-rights-mainstreaming-strategy-analysis-responses.pdf","source_class":"GOVERNMENT_OR_REGULATOR","claims_supported":["The official analysis states that all themes, including views expressed by very small numbers, are covered.","It expressly says rare views are not given less weight merely because more common comments exist."]}],"closest_analogue":"Mahajan's participatory-provenance framework for auditing representational loss in AI-mediated public-consultation summaries.","overlap":"Both audit compressed consultation records against source submissions, focus on omitted minority or dissenting perspectives, disaggregate exclusion patterns, and use the audit to improve an official summary rather than merely score its prose quality.","remaining_difference":"The hypothesis uses a low-dimensional theme basis, respondent-group reconstruction residuals, and consequence-weighted restoration thresholds; participatory provenance uses semantic coverage, transport, and causal diagnostics. That is a measurable implementation difference, but it does not preserve a distinct problem, workflow, causal lever, or intended outcome.","classification":"OBVIOUS_COLLISION","disposition":"REJECT","rationale":"A direct 2026 analogue already formalizes and tests auditing official consultation summaries for lost dissenting and semantically isolated submissions. Substituting reconstruction residuals for semantic-coverage diagnostics is too narrow for this elimination screen."},{"hypothesis_id":"H3","search_queries":["international crisis early warning spectral analysis eigenvalues escalation event data","conflict early warning dynamic systems eigenvalue stability political violence","ICEWS escalation forecasting hidden Markov crisis event counts","international crisis event data early warning indicators official"],"sources":[{"source_id":"H3-S1","title":"Early ViEWS: A prototype for a political Violence Early-Warning System","publisher":"Violence & Impacts Early-Warning System","url":"https://viewsforecasting.org/publications/early-views-a-prototype-for-a-political-violence-early-warning-system/","source_class":"PRIMARY_RESEARCH","claims_supported":["ViEWS combines slowly changing structural risks with quickly emerging news-derived triggers.","Its core forecast is an ensemble of static and dynamic nonlinear models evaluated with proper scoring rules.","The prototype forecasts the timing and location of political violence rather than a joint spectral instability within an already active interstate crisis."]},{"source_id":"H3-S2","title":"Introducing the ICBe Dataset: Very High Recall and Precision Event Extraction from Narratives about International Crises","publisher":"arXiv","url":"https://arxiv.org/abs/2202.07081","source_class":"PRIMARY_RESEARCH","claims_supported":["ICBe supplies structured, sequenced events across hundreds of historical international crises.","Its ontology includes 117 behaviors and is designed for statistical analysis of escalation and de-escalation sequences.","The authors warn that existing event systems can omit information needed to reconstruct historical episodes."]},{"source_id":"H3-S3","title":"EU Conflict Early Warning System fact sheet","publisher":"European External Action Service","url":"https://www.eeas.europa.eu/sites/default/files/documents/Factsheet%20-%20EWS.pdf","source_class":"OFFICIAL_GUIDANCE","claims_supported":["The EU system combines quantitative and qualitative conflict-risk evidence to support preventive action before escalation.","Its quantitative component models probability and intensity over a horizon of up to four years using structural indicators.","The official cycle includes scanning, prioritisation, shared assessment, monitoring, and follow-up."]}],"closest_analogue":"ViEWS, an operational research system combining dynamic models and event-derived triggers to forecast political violence.","overlap":"Both construct multivariate political-risk states, update warnings from temporal data, evaluate forecasts out of sample, and aim to trigger preventive action before visible escalation culminates in violence.","remaining_difference":"The hypothesis targets crisis-day transitions inside active interstate crises and alarms on a locally growing joint eigenmode spanning rhetoric, mobilization, sanctions, and incidents. ViEWS and the EU system forecast violence risk with ensembles or structural indicators, not a declared local-stability boundary. ICBe makes a historical test feasible. The distinction is testable against matched univariate thresholds and existing dynamic forecast baselines.","classification":"POSSIBLE_DISTINCTION","disposition":"ADVANCE","rationale":"Conflict early warning is crowded, but the retrieved analogues do not collapse the specific local spectral-stability alert into an existing indicator, ensemble, or hidden-state forecast."},{"hypothesis_id":"H4","search_queries":["public sector audit targeting network centrality referral agencies accountability","oversight complaint referral network eigenvector centrality audit selection","government audit risk assessment complaint volume agency referrals official guidance","accountability networks responsibility shifting public administration network analysis"],"sources":[{"source_id":"H4-S1","title":"The Anatomy of Blame: A Network Analysis of Strategic Responsibility-Shifting After a Systemic Disaster","publisher":"arXiv","url":"https://arxiv.org/abs/2510.21681","source_class":"PRIMARY_RESEARCH","claims_supported":["The study represents accusations among public, private, and regulatory actors as a directed network.","It finds reciprocal and clustered responsibility-shifting rather than isolated one-to-one assignments.","It demonstrates that accountability diffusion can be treated as an organizational network phenomenon."]},{"source_id":"H4-S2","title":"Network Interventions: Applying Network Science for Pragmatic Action in Public Administration and Policy","publisher":"arXiv","url":"https://arxiv.org/abs/2109.08197","source_class":"PRIMARY_RESEARCH","claims_supported":["The paper defines network interventions as using network data to identify strategies for behavior change, performance improvement, or other desired outcomes.","It explicitly extends network-intervention strategies to public-sector interorganizational and governance networks.","It notes that public-administration network research has been more descriptive and inferential than intervention-oriented."]},{"source_id":"H4-S3","title":"Audits and Evaluations FAQs","publisher":"Office of Inspector General, U.S. Department of Commerce","url":"https://www.oig.doc.gov/about/faqs/faqs-audits-evaluations/","source_class":"GOVERNMENT_OR_REGULATOR","claims_supported":["The OIG selects audit work through annual program risk assessment, legal requirements, requests, priorities, prior findings, and other identified risks.","The official process does not describe spectral centrality or referral-network perturbation as an audit-selection rule."]},{"source_id":"H4-S4","title":"Complaints Stage 2: Assess complaint using risk filter","publisher":"UK Health and Safety Executive","url":"https://www.hse.gov.uk/foi/internalops/og/ogprocedures/complaints/assess.htm","source_class":"OFFICIAL_GUIDANCE","claims_supported":["Complaints are prioritized through a risk filter, local factors, enforcement history, and escalation rules.","The procedure records some referrals and handoffs but evaluates complaints and dutyholders primarily through case-level risk categories."]}],"closest_analogue":"Siciliano and Whetsell's framework for purposeful network interventions in public-sector interorganizational and governance networks.","overlap":"The network-intervention framework already covers using governance-network data to select nodes or structural changes for public value, while responsibility-shifting research shows that accountability can circulate through directed organizational networks. Official oversight processes already use risk-based targeting and referral information.","remaining_difference":"The hypothesis specifically builds a longitudinal responsibility-transfer operator from complaint referrals, ranks modal participation, simulates node or reporting-rule perturbations, and validates audit targets on future unresolved or repeatedly transferred cases. The retrieved work does not show that exact audit-selection pipeline, and its incremental value over degree, volume, and ordinary risk filters is directly testable.","classification":"POSSIBLE_DISTINCTION","disposition":"ADVANCE","rationale":"The conceptual ingredients exist separately, but the searched sources leave a concrete operational and empirical distinction: spectral intervention on responsibility-transfer records for audit targeting."},{"hypothesis_id":"H5","search_queries":["political forecasting model validation regime shift structural break scenario model","institutional reform scenario model post shock validation political science","spectral gap mode drift model validity monitoring reduced order model governance","Federal Reserve SR 11-7 ongoing monitoring model risk management model performance official"],"sources":[{"source_id":"H5-S1","title":"Supervisory Guidance on Model Risk Management","publisher":"Board of Governors of the Federal Reserve System, Office of the Comptroller of the Currency, and Federal Deposit Insurance Corporation","url":"https://www.federalreserve.gov/frrs/guidance/supervisory-guidance-on-model-risk-management.htm","source_class":"OFFICIAL_GUIDANCE","claims_supported":["The guidance requires model validation, outcome analysis, ongoing monitoring, explicit limitations, and governance controls.","It states that changing data relevance or market conditions can make a model perform unexpectedly.","Meaningful performance deviation can warrant limits on use, adjustment, recalibration, or redevelopment."]},{"source_id":"H5-S2","title":"Detecting and Predicting Forecast Breakdowns","publisher":"The Review of Economic Studies / RePEc","url":"https://ideas.repec.org/a/oup/restud/v76y2009i2p669-705.html","source_class":"PRIMARY_RESEARCH","claims_supported":["The paper formalizes forecast breakdown as materially worse out-of-sample than in-sample performance.","It links forecast breakdowns to instability in the data-generating process and relates breakdown tests to structural-break tests."]},{"source_id":"H5-S3","title":"Forecasting with Breaks","publisher":"Handbook of Economic Forecasting / Elsevier","url":"https://www.sciencedirect.com/science/article/pii/S1574070605010128","source_class":"AUTHORITATIVE_SECONDARY","claims_supported":["Structural breaks create parameter instability and can undermine forecasts.","Consequences depend on the type of break and model, motivating break detection and forecast evaluation rather than calendar-only review."]},{"source_id":"H5-S4","title":"New institutionalism, critical junctures and post-crisis policy reform","publisher":"Australian Journal of Political Science","url":"https://www.tandfonline.com/doi/full/10.1080/10361146.2017.1409335","source_class":"PRIMARY_RESEARCH","claims_supported":["Political-science research characterizes crises as shocks that can loosen path dependencies and produce sudden institutional change.","This supports the premise that pre-shock institutional relationships may cease to describe post-crisis reform politics."]}],"closest_analogue":"Established model-risk governance combining structural-break or forecast-breakdown detection with ongoing performance monitoring and restrictions, recalibration, or redevelopment when validity fails.","overlap":"Both treat models as bounded simplifications, monitor whether changing conditions invalidate them, compare outputs with realized outcomes, document limitations, and restrict or revise model use after threshold failures rather than waiting for scheduled review.","remaining_difference":"The hypothesis uses spectral-gap shrinkage, retained-mode rotation, and reconstruction residuals as the trigger for suspending a reduced political-institution scenario model. The sources use general performance, outcome, and structural-break diagnostics rather than those modal statistics or that political workflow.","classification":"OBVIOUS_COLLISION","disposition":"REJECT","rationale":"The central governance claim—monitor a model for regime-sensitive validity failure and restrict, recalibrate, or redevelop it when thresholds fail—is already explicit in mature guidance and forecast-breakdown research. Modal metrics and a political-reform application are implementation choices, not a sufficient distinction at this shallow screen."}],"nominated_ids":["H1","H3","H4"],"replenishment_recommended":false}