{"schema_version":1,"experiment_id":"eoa_inverse_innovation_exp04_retrieval_first_paired20_20260802","cell_id":"invariant_mode_decomposition_design__medicine_healthcare","considered_ids":["H1","H2"],"selected_id":null,"dispositions":[{"hypothesis_id":"H1","lane":"COLLISION","reason":"Substantial collision with an issued bifurcation-based deterioration patent, locally linear physiological-mode research, eigenvalue-based critical-transition work, and an existing multivariate ward-alert product; the remaining operator-eigenmode-and-abstention bundle is too narrow and implementation-specific to support selection."},{"hypothesis_id":"H2","lane":"COLLISION","reason":"Substantial collision with decades of multivariable dialysis biofeedback, patient-specific adaptive control, online-updated model-predictive control, and robust ultrafiltration design; the surviving SVD and reconstruction-gate distinction is an algorithmic formulation rather than a differentiated treatment lever."}],"selection_reason":"Neither hypothesis presents a sufficiently material, repairable residual treatment claim after the critics' direct-source collisions; both are routed to the collision lane and no finalist is selected."}