id	author	title	date	pages	extension	mime	words	sentences	flesch	summary	cache	txt
work_ms5ojw7dunfmhanpja5c5cgjse	Nikolaos Pappas	GILE: A Generalized Input-Label Embedding for Text Classification	2019	16	.pdf	application/pdf	10703	1205	63	GILE: A Generalized Input-Label Embedding for Text Classification generalizes over previous such models, addresses their limitations, and does not compromise performance on seen labels. (iii) they are outperformed on seen labels by classification baselines trained with cross-entropy loss non-linear input-label embedding with controllable capacity and a joint-space-dependent classification unit which is trained with cross-entropy (ii) We propose a novel joint input-label embedding with flexible parametrization which generalizes over the previous such models and Joint input-output embedding models can generalize from seen to unseen labels because the parameters of the label encoder are shared. on the generalized input-label embedding outperforms previous models with a typical output layer Table 3: Full-resource classification results on general (upper half) and specific (lower half) labels using monolingual and bilingual models with DENSE encoders on English as target (left) and the auxiliary language as target	./cache/work_ms5ojw7dunfmhanpja5c5cgjse.pdf	./txt/work_ms5ojw7dunfmhanpja5c5cgjse.txt