Profile cover photo
Profile photo
Jon Orwant
1,294 followers -
Engineering Director, Google Research
Engineering Director, Google Research

1,294 followers
About
Jon's posts

Post has attachment
Let's dispel with this fiction that AlphaGo doesn't know what it's doing. AlphaGo knows exactly what it's doing.

http://www.theverge.com/2016/3/12/11210650/alphago-deepmind-go-match-3-result

Post has shared content
An amazing achievement. Congrats to the Deep Mind team!
Can a computer master the ancient game of Go? With more possible board configurations than the number of atoms in the universe, experts have long thought it to be impossible. But a system developed by researchers at Google DeepMind, called AlphaGo, has done it.

Combining a state-of-the-art Monte-Carlo tree search with two deep neural networks, AlphaGo not only has defeated the top existing Go programs at the forefront of A.I. research, but also has defeated reigning 3-time European Go champion Fan Hui. 

The next challenge will be to play the top Go player in the world over the last decade, Lee Sedol, this March in Seoul, South Korea. To learn more, visit the Google Research blog, linked below.

Post has attachment

Post has shared content

Post has attachment
Best lede ever: "A former meerkat expert at London Zoo has been ordered to pay compensation to a monkey handler she attacked with a wine glass in a love spat over a llama-keeper."

My 12 year old daughter starts 7th grade today, in a new school. It begins with the principal greeting all students in the auditorium. 

I've encouraged my daughter to sit in the back. When the principal begins speaking, she is to stand up and yell "I VOLUNTEER AS TRIBUTE!"

Junior high and Hunger Games have a lot in common.

Post has shared content
All of these images were computer generated!

For the last few weeks, Googlers have been obsessed with an internal visualization tool that Alexander Mordvintsev in our Zurich office created to help us visually understand some of the things happening inside our deep neural networks for computer vision.  The tool essentially starts with an image, runs the model forwards and backwards, and then makes adjustments to the starting image in weird and magnificent ways.  

In the same way that when you are staring at clouds, and you can convince yourself that some part of the cloud looks like a head, maybe with some ears, and then your mind starts to reinforce that opinion, by seeing even more parts that fit that story ("wow, now I even see arms and a leg!"), the optimization process works in a similar manner, reinforcing what it thinks it is seeing.  Since the model is very deep, we can tap into it at various levels and get all kinds of remarkable effects.

Alexander, +Christopher Olah, and Mike Tyka wrote up a very nice blog post describing how this works:

http://googleresearch.blogspot.com/2015/06/inceptionism-going-deeper-into-neural.html

There's also a bigger album of more of these pictures linked from the blog post:

https://goo.gl/photos/fFcivHZ2CDhqCkZdA

I just picked a few of my favorites here.
PhotoPhotoPhotoPhotoPhoto
2015-06-17
18 Photos - View album

Post has attachment
Photo

Post has attachment
"And so a trend for the worse is grimly established."

"The suggestion becomes dark fact with one step farther into the past."

This is writing that leaps off the page and into seventh grade English class. Not to worry, though, there's a cheery cartoon on the next page.
Photo

Post has attachment
In 1951 there were still enough horse-drawn vehicles on the road for that to be a thing.
Photo
Wait while more posts are being loaded