3 ms·
> Mapping images to categories, speech to text, text to categories, go boards to good moves, and the like, is extremely useful, and cannot be done as well with
by YeGoblynQueenne 11y ago
> Mapping images to categories, speech to text, text to categories, go boards to good moves, and the like, is extremely useful, and cannot be done as well with other methods.
[citation needed]
- nl 11y agoWhich part do you need the citation for? Images to Categories: http://image-net.org/challenges/LSVRC/2015/results http://image-net.org/challenges/LSVRC/2015/results Speech to Text: https://en.wikipedia.org/wiki/Speech_recognition#Deep_Feedforward_and_Recurrent_Neural_Networks https://en.wikipedia.org/wiki/Speech_recognition#Deep_Feedfo... gives a good set of citations of different benchmarks text to categories: https://papers.nips.cc/paper/5782-character-level-convolutional-networks-for-text-classification.pdf https://papers.nips.cc/paper/5782-character-level-convolutio... (Note that DL methods only improve on manual feature engineering on the more difficult tasks here) go boards to good moves: The recent DeepMind games?