Monthly:August 2016

Applying machine learning to radiotherapy planning for head & neck cancer

Were excited to announce a new research partnership with the Radiotherapy Department at University College London Hospitals NHS Foundation Trust, which provides world-leading cancer treatment.1 in 75 men and 1 in 150 women will be diagnosed with oral cancer during their lifetime, and oral cavity cancer has risen by 92% since the 1970s.Read More

Applying machine learning to radiotherapy planning for head & neck cancer

Were excited to announce a new research partnership with the Radiotherapy Department at University College London Hospitals NHS Foundation Trust, which provides world-leading cancer treatment.1 in 75 men and 1 in 150 women will be diagnosed with oral cancer during their lifetime, and oral cavity cancer has risen by 92% since the 1970s.Read More

Google Cast is now built in to Chrome

Two years ago we launched Google Cast, an extension for Chrome that allowed you to show content from your favorite websites on the best screen in your home—your TV—using Chromecast.  Whether watching your favorite show on Netflix, or sharing a photo slideshow at a family gathering, Google Cast makes it easy to extend the web to the big screen.  

Since we launched Chromecast, we’ve been working to make casting even better. We’ve launched new Cast devices like Chromecast Audio and partnered with other manufacturers to make Cast-enabled TVs and speakers.  We’ve also made significant improvements in quality, with the vast majority of casting sessions now in HD.

Casting from Chrome has become incredibly popular: In the past month alone, people have casted more than 38 million times from Chrome, watching and listening to more than 50 million hours of content.

Today we’re happy to announce that Google Cast is now built fully into Chrome, and anyone can now Cast without having to install or configure anything.  When you’re on websites that are integrated with Cast, you’ll see the Cast icon appear when you’re on the same network as a Cast device—and with a couple of clicks you can view your content on your TV or listen to music on your speakers:

google cast in chrome 1.png

You can also view websites that are not integrated with Cast on your TV by selecting the “Cast…” menu item from the Chrome menu:

google cast in chrome 2.png

And you can now cast to even more places—including sharing with participants on a video call in Google Hangouts and the Cast for Education app, which we announced recently.  Now you can share your presentation with your coworkers during a Hangout or to your peers in the classroom.

With Google Cast and Chrome, you can share what you love to watch with those around you. We’re excited to make this available to everyone.  To get the new built-in Cast functionality all you need to do is make sure you have the latest version of Chrome.

Read More

Decoupled Neural Interfaces Using Synthetic Gradients

Neural networks are the workhorse of many of the algorithms developed at DeepMind. For example, AlphaGo uses convolutional neural networks to evaluate board positions in the game of Go and DQN and Deep Reinforcement Learning algorithms use neural networks to choose actions to play at super-human level on video games.This post introduces some of our latest research in progressing the capabilities and training procedures of neural networks called Decoupled Neural Interfaces using Synthetic Gradients. This work gives us a way to allow neural networks to communicate, to learn to send messages between themselves, in a decoupled, scalable manner paving the way for multiple neural networks to communicate with each other or improving the long term temporal dependency of recurrent networks. This is achieved by using a model to approximate error gradients, rather than by computing error gradients explicitly with backpropagation. The rest of this post assumes some familiarity with neural networks and how to train them. If youre new to this area we highly recommend Nando de Freitas lecture series on Youtube on deep learning and neural networks.Neural networks and the problem of lockingIf you consider any layer or module in a neural network, it can only be updated once all the subsequent modules of the network have been executed, and gradients have been backpropagated to it.Read More