You can't use the data from one learning process to enhance the process of learning either a related or an unrelated task I beg to differ sir: http://nlp.stanf

You can't use the data from one learning process to enhance the process of learning either a related or an unrelated task I beg to differ sir: http://nlp.stanford.edu/pubs/SocherGanjooManningNg_NIPS2013.pdf http://nlp.stanford.edu/pubs/SocherGanjooManningNg_NIPS2013.... This paper demonstrates something called "0 shot learning" where you can actually infer the correct label of an unseen image based on similarity among representations learned in a separate NLP task. For instance, it can label an image "tiger" even if it has not seen tigers but has only learned about the word (and inferred its relation to cat, an image it has seen) from reading text. It's not intelligent, not even close. But it's an awfully strange emergent phenomenon these concepts are demonstrating. Exciting stuff, I think.