Profile cover photo
Profile photo
Peter Liu
Machine teacher
Machine teacher
About
Peter's posts

Post has shared content
Wes Anderson vibes in this Switzerland hotel by Johannes Hulsch: http://goo.gl/oL2Cu5 #travel
Photo

Post has shared content
Inspiring words from an NFL football player who just published a(nother) mathematical research paper. One neat quote:

I have no desire to try to accumulate $10 million in the bank; I already have more money in my bank account than I know what to do with. I drive a used hatchback Nissan Versa and live on less than $25k a year. It's not because I'm frugal or trying to save for some big purchase, it's because the things I love the most in this world (reading math, doing research, playing chess) are very, very inexpensive.

Post has shared content
nice, saw on chromecast
紅富士山と一本松
Mt. Fuji in pink and solitude pine tree early in the morning

昨日の投稿の別バージョンです。こちらの方が撮影時間が少し早く、富士山が綺麗なピンク色になっています。

#cooljapan #NatureJP #fujisan #mtfuji #富士山 #一本松 #landscape #Canon #EOS6D
Photo

Post has shared content
Do NOT buy a +Nest smoke alarm. They false alarm and are unhushable pieces of crap. Please spread the word.

Watch this horror video of me trying to quiet mine.

You can stop or mute this video if it's annoying, but you cannot stop an actual Nest protect.

This went off in my house all day, annoying my neighbors.

When I got the Android notification that my house was burning down I immediately assumed it was false, since my Nests had already cried wolf before.  I also checked video cameras and saw my house wasn't actually burning down, so I stayed at work. My poor neighbors, though.

Disclaimer: I am a Google employee. I paid for these myself. So I speak as myself.

Post has shared content
The End of the World is Coming by Fabio Marchini: http://goo.gl/wGzsta
Photo

Post has shared content
Dazzling Make-Up Transforms Human Face From 3D Into 2D. Artist: Valeriya Kutsan
PhotoPhotoPhotoPhotoPhoto
Make-up artist Valeriya Kutsan
8 Photos - View album

Post has attachment
This is Dallas at night
Photo

Post has shared content
Baidu Research just attained the best computer vision ImageNet classification result! 5.98% error (vs. GoogLeNet's 6.66%). The key to this was our multi-GPU deep learning infrastructure, which by using a mix of model-parallelism and data-parallelism, allows us to train our model 24.7x faster than using only a single GPU. This scale also allows us to use higher-resolution images, and absorb more (synthetic) training data. Paper here: bit.ly/deepimage
Photo

Post has shared content
Who says there can't be innovation in sports? From Germany, meet Headis, a combination of...well, you'll have to see it yourself to believe it. And these players are GOOD.

Post has shared content
Wait while more posts are being loaded