This AI experiment generates realistic handwriting: Bring on the identity stealing robots

Screenshot 2015-07-24 09.14.41
This isn’t new – it’s from 2013 – but it hasn’t got nearly enough attention. Alex Graves at Toronto University built this handwriting generation demo that uses recurrent neural networks to produce convincing cursive script. Here’s the explanation of how it works from the abstract of Graves’ paper on the project: This paper shows how Long Short-term Memory recurrent neural networks can be used to generate complex sequences with long-range structure, simply by predicting one data point at a time. The approach is demonstrated for text (where the data are discrete) and online handwriting (where the data are real-valued). It is then…

This story continues at The Next Web








from The Next Web » Apps http://ift.tt/1fr9jVB

0 commentaires:

Enregistrer un commentaire