G-RELIC uses LLMs and graph-based mapping to improve consistency in clinical document relexicalization. The method ensures one-to-one entity correspondence and utilizes a deterministic algorithm to maintain temporal consistency in longitudinal records.
HOW THIS AFFECTS YOU
●
builderYou can generate higher-fidelity synthetic clinical datasets that maintain relational and temporal coherence.
●
healthThis improves the utility of de-identified clinical data for downstream longitudinal scientific research.