GapDream: Active Synthetic Evidence Acquisition for Long-Tailed Medical Imaging

Synthetic medical data are commonly generated from manually chosen prompts, inverse-frequency sampling, or fixed subgroup-balancing rules. These strategies increase sample count but do not answer a more consequential question: which missing observations would most improve the current model? We introduce G ap -D ream , a closed-loop framework that treats a medical image generator as a queryable source of candidate evidence rather than an unlimited source of augmentation. At each round, G ap -D ream maps model failures into an evidence-gap score that combines scarcity, predictive uncertainty, local representation density, and gradient novelty. A query policy translates high-value gaps into structured clinical requests, a modality-specific generator proposes candidate image–text pairs, and an independent selector retains only diagnostically consistent, diverse, non-duplicative candidates. The downstream model is then updated and the gap map recomputed. The central objective is therefore not image realism in isolation, but reduction of measurable evidence gaps under a fixed synthetic budget. Across a multi-modality protocol spanning long-tailed chest radiography, dermoscopy, clinical dermatology, and histopathology, G ap D ream reaches stronger held-out real-data performance with substantially fewer generated samples than random, class-balanced, uncertainty-only, and static targeted augmentation. Results show gains in CXR-LT tail mAP, HAM10000 macro-AUROC, Camelyon17 out-of-domain accuracy, and worst-group dermatology accuracy while reducing redundant synthetic samples. G ap D ream reframes medical generation as active evidence acquisition: generate the examples the learner needs, not simply more examples of what is rare.