Evaluating the Utility of Hand-crafted Features in Sequence Labelling

Open paper

Benchmark	Model	Rank	Results
named-entity-recognition-ner-on-conll-2003	Neural-CRF+AE	#41	F1: 92.29
named-entity-recognition-ner-on-conll-2003	CRF + AutoEncoder	#48	F1: 91.87