{"schema_version":1,"assessment_id":"eoa_inverse_innovation_exp03_opportunity320_20260801","source_experiment_id":"eoa_inverse_innovation_exp03_full320_20260801","cell_id":"invariant_mode_decomposition_design__material_culture_museum_studies","archetype_slug":"invariant_mode_decomposition_design","domain_slug":"material_culture_museum_studies","title":"Modal disruption of recurring museum-catalog classifications","opportunity_summary":"Test whether persistent coupled workflow patterns explain the recurrence and downstream exposure of classifications that authorized reviewers have identified as contested or obsolete, and whether reversible mode-targeted workflow changes outperform ordinary record-level cleanup without compromising provenance, plural description, discoverability, or human authority.","adopter_authorizer":"A museum's catalog or data-governance authority and accountable collection leadership, jointly with appropriately authorized representatives of affected communities for culturally sensitive classifications.","scores":{"meaningful_impact":{"score":3,"rationale":"If recurrence is material, weakening propagation paths could reduce repeated harmful exposure and cleanup effort while preserving provenance. The packet provides no observed prevalence, downstream reach, or affected collection scale, so aggregate impact remains uncertain."},"stakeholder_pull":{"score":2,"rationale":"Catalogers, curators, collections managers, affected communities, researchers, and visitors have plausible interests in the stated outcome, but the sealed candidate contains no expressed demand, adoption commitment, complaint volume, or evidence that recurrence is a current priority."},"incremental_advantage":{"score":3,"rationale":"The proposal makes a comparative claim against matched record-level cleanup and identifies a strong non-spectral dependency-audit rival. Coordinated mode targeting could address coupled propagation missed by isolated corrections, but no held-out evidence shows an advantage over explicit-rule, single-source, or process-tracing approaches."},"distinctiveness_plausibility":{"score":2,"rationale":"The combination of modal propagation analysis, reversible targeting, residual governance, and shared community authority is concretely specified, but prior art is explicitly unsearched. Distinctiveness relative to existing catalog remediation and workflow-analysis methods therefore cannot be established closed-book."},"technical_implementability":{"score":3,"rationale":"The read-only analysis, audit-window comparison, sensitivity testing, snapshots, and reversible pilot are technically bounded. Feasibility depends on sufficiently complete transition logs, consistent semantic coding, stationarity, conditioning, diagnostic stability, and intervention mapping, none of which is demonstrated."},"adoption_authority_feasibility":{"score":3,"rationale":"The proposal correctly locates authority with museum governance and authorized affected-community representatives and excludes automated terminology decisions. Feasibility is tempered by unspecified representative selection, joint approval, consultation burden, and potential disagreement over culturally sensitive classifications."},"evidence_readiness":{"score":3,"rationale":"Observable states, baselines, rivals, falsifiers, held-out comparisons, residual tests, and rollback rules are specified well enough to design a study. No transition data, recurrence measurements, operator estimates, or verified manipulation evidence is supplied."},"safety_net_benefit":{"score":4,"rationale":"Read-only analysis, an internal reversible pilot, provenance preservation, metadata snapshots, explicit prohibited uses, affected-party objection rights, and multiple halt criteria provide a strong safety structure. Residual harms, reification, authority inequity, and discoverability loss still require empirical and human review."},"scalability":{"score":2,"rationale":"The approach may be reusable as an analytical pattern, but each collection can have different vocabularies, authority structures, software, audit completeness, policies, and community-governance requirements. Mode drift and collection-specific validity windows limit straightforward replication."}},"score_confidence":"MODERATE","costs":{"first_evidence":{"band_2026_usd":"50K_TO_250K","scope":"Pre-register and conduct a read-only retrospective study on one collection using separated audit windows; code authorized classifications and workflow touchpoints; compare the modal model with explicit-rule, single-source, non-spectral dependency-audit, and ordinary-cleanup predictors on held-out recurrence and exposure.","confidence":"LOW","assumptions":["Usable metadata snapshots and audit logs already exist and can be accessed internally.","The collection is bounded enough for manual coding and reviewer adjudication.","Costs include catalog, data, evaluation, governance, and affected-community review labor.","No production workflow or public record is changed."]},"initial_deployment_startup":{"band_2026_usd":"50K_TO_250K","scope":"Prepare one approved eight-week internal-only reversible pilot, including workflow mapping, intervention configuration, matched baseline construction, snapshots, approvals, staff preparation, monitoring, and rollback testing.","confidence":"LOW","assumptions":["The read-only study first establishes a stable and decision-relevant mode.","Changes can be configured without replacing core catalog systems.","Only one collection and one targeted mode are included.","Required museum and affected-community approvals can be obtained without unusually extensive coordination."]},"operational_launch":{"band_2026_usd":"250K_TO_1M","scope":"Move a validated method into governed operation for the named collection, including production integration, reviewed record or rule changes, provenance controls, access testing, community governance, staff training, monitoring, and independent evaluation.","confidence":"LOW","assumptions":["Launch remains limited to one institution and collection.","Existing catalog, search, synchronization, and exhibition-selection systems require multiple integrations but not wholesale replacement.","Public-facing changes occur only after required human review.","The band includes compliance, partner coordination, evaluation, and contingency work."]},"annual_recurring":{"band_2026_usd":"50K_TO_250K","scope":"Maintain the named-collection process through drift and residual monitoring, periodic re-estimation, audit-log upkeep, software maintenance, staff review, affected-community participation, discoverability checks, and incident response.","confidence":"LOW","assumptions":["The system remains limited to a small number of monitored modes in one collection.","Human review and community governance remain mandatory.","Material policy or software changes trigger revalidation rather than automatic continuation.","No major platform migration is included."]}},"research_burden":"HIGH","earliest_credible_horizon":"3_TO_12_MONTHS","pipeline_gates":{"recognizable_externally_supportable_problem":{"status":"UNCERTAIN","reason":"Repeated recurrence after correction is clearly recognizable and operationalized, but it is labeled a hypothesis and the packet supplies no observations showing coupled recurrence, prevalence, or downstream exposure in the proposed collection."},"identifiable_adopter_or_authorizer":{"status":"YES","reason":"The candidate identifies museum catalog or data-governance authority and accountable collection leadership, with authorized affected-community representatives sharing authority for culturally sensitive classifications."},"distinct_testable_incremental_claim":{"status":"YES","reason":"It claims that targeting a stable propagation mode will reduce held-out recurrence and exposure more than matched record-level cleanup, while naming non-spectral dependency auditing, explicit rules, and single-source explanations as comparisons."},"bounded_next_evidence_step":{"status":"YES","reason":"A pre-registered, read-only analysis on one collection can compare models across separated audit windows and reject the proposal for weak stability, structured residuals, or no held-out gain without changing live records."},"no_unresolved_safety_or_authority_stop":{"status":"UNCERTAIN","reason":"The proposal supplies strong exclusions, joint authority, halt conditions, snapshots, and rollback, but actual affected-community representation, approval, consultation burden, and acceptable cross-mode or discoverability effects have not been established."},"implementation_cost_scope_and_range":{"status":"UNCERTAIN","reason":"The one-collection and eight-week-pilot scopes are bounded enough for broad resource bands, but collection size, audit-log quality, system integrations, review volume, and coordination requirements are unspecified and could materially change cost."}},"blocking_evidence":["Observed evidence that authorized classifications recur across workflow cycles in coupled patterns after controlling for one-way imports, explicit rules, and dominant single fields.","Adequate audit-log completeness, semantic coding reliability, stationarity, conditioning, spectral separation, mode stability, and acceptably unstructured held-out residuals across separated windows.","Comparative predictive and intervention advantage over explicit-rule, single-source, non-spectral dependency-audit, and ordinary record-level-cleanup approaches.","Evidence that a verified mode-targeted manipulation reduces held-out recurrence and exposure without unacceptable cross-mode, provenance, plural-description, or discoverability effects.","A workable joint-authority process with appropriately authorized affected-community representatives and an acceptable consultation burden.","Bounded prior-art findings clarifying whether the proposed composition is distinct from existing catalog-remediation and workflow-analysis methods.","Collection-specific integration, staffing, review, and monitoring requirements sufficient to refine the broad cost ranges."],"next_evidence_step":"Pre-register a read-only retrospective study on one named collection using separated development and held-out audit windows. Compare the modal model against explicit-rule, dominant-source, non-spectral dependency-audit, and ordinary-cleanup baselines for recurrence and downstream-exposure prediction. Stop before any live change if coupled recurrence is absent, a simpler model performs equivalently or better, conditioning or mode stability fails, or held-out residuals remain structured above the pre-set tolerance.","research_questions":["Do corrected classifications demonstrably recur across cycles, and what share is explained by explicit rules, one-way imports, or a single upstream authority file?","Are coded workflow states reliable enough to estimate a sufficiently stationary, well-conditioned, and diagnostically stable operator across separated windows?","Does the modal model improve held-out recurrence and exposure prediction over the non-spectral dependency audit and simpler baselines?","Can sensitivity estimates identify a reversible workflow change whose effects are distinguishable from association or changing policy and staffing?","What residual, cross-mode, provenance, plural-description, and discoverability thresholds will authorized reviewers consider acceptable?","Who can legitimately represent affected communities for this collection, and what approval, objection, compensation, and withdrawal procedures are required?","Which elements, if any, are distinct from existing museum catalog remediation, metadata-governance, and workflow-analysis approaches?","What collection size, system interfaces, log quality, staff effort, and review volume determine implementation cost? "],"recommendation":"PARTNERED_RESEARCH","uncertainty_constraints":["Problem prevalence, downstream reach, and stakeholder demand are not measured.","Prior art, world novelty, market size, and realized impact are outside the sealed evidence.","Operator estimability, stationarity, diagonalizability, conditioning, spectral gap, and causal adequacy are untested.","Semantic coding may manufacture apparent modes or suppress consequential residual patterns.","Policy, staffing, incentives, software, or one authority file may explain recurrence more simply.","Community authorization, representativeness, consultation burden, and acceptable terminology cannot be inferred mathematically.","Cost bands are resource-equivalent planning ranges, not estimates based on a specified institution, collection size, or system architecture.","Cross-collection scalability is constrained by local workflows, vocabularies, governance, and validity windows."],"closed_book_prior_art_boundary":"Prior-art status is explicitly UNSEARCHED. This assessment makes no claim about novelty, prevalence, existing implementations, market size, realized impact, or comparative performance outside the sealed candidate; distinctiveness requires a separate bounded review."}