id	author	title	date	pages	extension	mime	words	sentences	flesch	summary	cache	txt
orr-bootleg-2020			2020-10-26	25	.pdf	application/pdf	14458	1211	63	in a knowledge base, is how to disambiguate entities that appear rarely in the training data, termed tail Humans use subtle reasoning patterns based on knowledge of entity facts, relations, and types to transfer to other non-disambiguation tasks that require entity-based knowledge: we set a new state-ofthe-art in the popular TACRED relation extraction task by 1.0 F1 points and demonstrate up to 8% shows F1 versus number of times an entity was seen in training data for a baseline NED model compared to 3 Bootleg encodes the entity, relation, and type signals as embedding inputs to a unseen entities compared to the two models which respectively use only type and only relation the discriminative entity and more general relation and type signals that are useful for disambiguation. Given an example, we run inference with the Bootleg model to disambiguate named entities and generate	./cache/orr-bootleg-2020.pdf	./txt/orr-bootleg-2020.txt