Re:Why do we use word embeddings in NLP?
by buckhx on Tuesday September 10, 2019 @10:31AM
(#20928616)
BERT & GPT-2 are state of the art, pre-trained models with transfer learning in mind.
Re:Why do we use word embeddings in NLP?
by superpermutat0r on Tuesday September 10, 2019 @10:17AM
(#20928440)
https://www.cv-foundation.org/openaccess/content_cvpr_worksh...
Here's a nice paper. But yes, almost any CNN works.
VGG-16 or VGG-19 seem to be used the most.
Re:Why do we use word embeddings in NLP?
by thanatropism on Tuesday September 10, 2019 @01:32PM
(#20930813)
Yeah, but if I understood it correctly the VGG-16 that comes with Keras is trained to give a probability distribution over 1000 interpretable labels. I was hoping for something that, like word2vec, embedded image data in low-ish dimension.
I know easy-to-use object recognition models, but for general clustering, metric learning, etc. tasks it would be useful to have an abstract embedding.