A comparison of few-shot and traditional named entity recognition models for medical text | IEEE Conference Publication | IEEE Xplore