OptimCLM: Optimizing clinical language models for predicting patient outcomes via knowledge distillation, pruning and quantization

Mohammad Junayed Hasan, Fuad Rahman, Nabeel Mohammed

Abstract: Highlights•The study integrated clinical outcome prediction into each stage of a patient's healthcare journey.•This aids healthcare systems by planning, optimizing resources and improving patient outcomes.•Clinical language models were compressed using ensemble-knowledge distillation, pruning, and quantization.•This framework achieved up to 22.88× compression and 28.7× speedup with minimal AUROC loss.•The ensemble outperformed state-of-the-art models on four major clinical outcome prediction tasks.