Post has attachment
@FredBouchery reframes @mo_selim_art 's video to illustrate the classic trade-off between having it good, or having it fast. (The third part of the trifecta, whether it was cheap, is already constrained by having a single illustrator of a particular skill.)

Original video: https://t.co/f0ZLRG2sjX (credit @mo_selim_art )
Add a comment...

I know backwards compatibility sometimes needs kludges, but somehow I think Microsoft could have made the naming less confusing...

"it looks wrong to launch the 32-bit version from a folder with 64 in the name, but that is correct.

To add to the confusion, the C:\Windows\Sys* folder names are dynamic. When you are in a 64 bit session, the 64-bit folder is named System32, the 32-bit folder is named SysWOW64, and the SysNative folder does not exist. When you are in a 32-bit session, the 64-bit folder is named SysNative, the 32-bit folder is named System32, and the SysWOW64 folder does not exist."
Add a comment...

Post has shared content
Very interesting approach!
Moving Beyond the Turing Test with the Allen AI Science Challenge

The field of artificial intelligence has made great strides recently, as in AlphaGo's victories in the game of Go over world champion South Korean Lee Sedol in March 2016 and top-ranked Chinese Go player Ke Jie in May 2017, leading to great optimism for the field. But are we really moving toward smarter machines, or are these successes restricted to certain classes of problems, leaving others untouched? In 2015, the Allen Institute for Artificial Intelligence (AI2) ran its first Allen AI Science Challenge, a competition to test machines on an ostensibly difficult task—answering eighth-grade science questions. Our motivations were to encourage the field to set its sights more broadly by exploring a problem that appears to require modeling, reasoning, language understanding, and commonsense knowledge in order to probe the state of the art while sowing the seeds for possible future breakthroughs....Our goal with the Allen AI Science Challenge was to operationalize one such test—answering science-exam questions. Clearly, the Science Challenge is not a full test of machine intelligence but does explore several capabilities strongly associated with intelligence—capabilities our machines need if they are to reliably perform the smart activities we desire of them in the future, including language understanding, reasoning, and use of common-sense knowledge. Doing well on the challenge appears to require significant advances in AI technology, making it a potentially powerful way to advance the field. Moreover, from a practical point of view, exams are accessible, measurable, understandable, and compelling. One of the most interesting and appealing aspects of science exams is their graduated and multifaceted nature; different questions explore different types of knowledge, varying substantially in difficulty, especially for a computer.
Add a comment...

Post has shared content
Think about forbidding users from choosing these common passwords on your site
Add a comment...

Post has shared content
Point 2 is a very interesting essay on what it is to be at engineer
A Googler wrote an (internal, since leaked) manifesto about gender and engineering a few days ago. If you haven't read it, I will say that you are not missing much. But if you've heard about this and are wondering what my response was, I just posted it publicly.

The intro of this also hints at some bigger news which I'll get into later. :)
Add a comment...

Post has shared content
Hmm, maybe +Tom C​ can use this his security/wildlife camera...
"I created an object detector that is able to recognize Racoons with relatively good results." This is with TensorFlow's new Object Detector API that you can use to train an object detector with your own dataset.
Add a comment...

Post has attachment
Add a comment...

Post has attachment
"There are only two hard problems in distributed systems: 2. Exactly-once delivery 1. Guaranteed order of messages 2. Exactly-once delivery" - +Mathias Verraes

Nice distributed turn on an old quip!
Add a comment...

Post has shared content
Interesting... Sounds worth looking into
"In following its externalization trend, Google has risen to this challenge by releasing many of its internal AI systems as public cloud APIs. Cloud Vision accepts any arbitrary image and catalogs objects and activities, OCRs text, recognizes the location depicted, estimates the emotion of human faces and even flags whether the image depicts violence. All with a single API call and with results returned in just a few seconds and infinitely scalable. Cloud Speech performs live speech to text in over 80 languages and, unlike legacy speech transcription systems, requires no training and is incredibly robust to noise. Cloud Natural Language accepts arbitrary text in English, Spanish and Japanese and outputs a robust dependency parse tree, recognizes key entities and even performs sentiment analysis. At Next '17, Google expanded this lineup with its latest tool, Cloud Video Intelligence, which takes a video and segments it into scenes and identifies the major topics and activities in each scene, allowing one to take a massive video archive and instantly index it to make it topically searchable."
Add a comment...

Post has shared content
Via +Yonatan Zunger​, who added:

"Movies have always loved having technicians magically "enhance" a blurry image, somehow turning pixels into clear pictures. Anyone who knows how computers work has always laughed at these, because there is literally no data beyond the pixels given; there's nothing there to enhance.

Except maybe there is. This program does something clever: it takes a pixelated image, and uses the fact that it knows it's looking at a human face, and what human faces look like, to turn each pixel into a 4x4 grid of its best guess of which colors would have to have been there to both be consistent with a face shape and with the average color it saw.

The fact that this works at all is pretty stunning, but take a look at the output below. On the right are the original pictures, at 32x32 resolution. On the left is what happens after they're reduced down to 8x8, the sort of thing you would get when a camera is at the limit of its resolution. In the middle is what their algorithm recovered.

The paper (available at the link) shows outputs for other kinds of image as well, e.g. "pictures of bedrooms." It clearly has to be trained afresh for each type of subject matter, and it's not yet clear how much its abilities scale from one kind of subject to a similar one, but it's quite impressive.

h/t +Kee Hinckley"
Algorithm that increases the resolution of images, synthesizing additional details as needed based on its understanding of what e.g. human faces should look like. Left is an 8x8 input, middle is the enhanced 32x32 version, right is what the 8x8 input originally looked like at 32x32 size.

Paper: https://arxiv.org/abs/1702.00783
Photo
Add a comment...
Wait while more posts are being loaded