Google has revealed its new neural network system for multi-tasking which is called MultiModal. The system is capable of learning how to detect objects in images, provide captions, recognize speech, translate between four pairs of languages as well as parse grammar and syntax, all at once.
At Google I/O, Google shared the results from its Quick, Draw! neural network testing. These results show how 15 million people drew the same set of objects.
The developers of FaceApp have released the Android version of their neural network-powered application. The app allows users to add a smile to someone in a photo who wasn't smiling, change their gender or see what they would look like if they were older or younger. The FaceApp can be downloaded...
Google's DeepMind neural network is great at making quick, reactive decisions, but it's not so great at making careful, complex choices.
However, Google is experimenting with a type of memory-augmented neural network that could eventually help you to navigate the subway.
Google has partnered with Carnegie Mellon and Cornell to create a new machine learning database that's filled with over 9 million photo and video assets which it claims can be used to train a neural network from scratch. All items in the database were tagged by computers and then verified by...
In the past, Google has done their translations one word at a time and this can sometimes cause errors. However, the company is using their neural networks to create a Google Neural Machine Translation system (GNMT). This system allows Google to translate entire sentences at once instead of one...
Google's Android security team feels that if they continue on the black vs. white path when it comes to malware then it won't get them to where they need to go.
Instead, Adrian Ludwig wants to take a page from credit card companies where they build 'risk profiles' for person, and then apply it...
AI and Google's neural networks can be used for a lot of different things and the Mountain View tech giant is now experimenting with poetry.
Google is said to have "inspired" their neural networks by giving the system two sentences from all the books it had read so far. Then they programmed it...
This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
By continuing to use this site, you are consenting to our use of cookies.