Conceptual

Pan-Infection Foundation Model with Knowledge Distillation for Host-Response Pathogen Diagnosis

A diagnostic modeling strategy that trains one broad host-response classifier on pooled blood-transcriptome data across many infections, then distills it into small pathogen-specific models. A student learns how a foundation-then-distill pipeline transfers diagnostic knowledge from a large teacher to deployable lightweight classifiers.