Google Unveils Neural Network With Ability To Determine Location of Any Image (technologyreview.com)

← Back to Stories (view on slashdot.org)

Google Unveils Neural Network With Ability To Determine Location of Any Image (technologyreview.com)

Posted by BeauHD on Saturday February 27, 2016 @11:28AM from the neural-networking dept.

schwit1 writes: Here's a tricky task. Pick a photograph from the web at random. Now try to work out where it was taken using only the image itself. If the image shows a famous building or landmark, such as the Eiffel Tower or Niagara Falls, the task is straightforward. But the job becomes significantly harder when the image lacks specific location cues or is taken indoors or shows a pet or food or some other detail. Nevertheless, humans are surprisingly good at this task. To help, they bring to bear all kinds of knowledge about the world such as the type and language of signs on display, the types of vegetation, architectural styles, the direction of traffic, and so on. Humans spend a lifetime picking up these kinds of geolocation cues. So it's easy to think that machines would struggle with this task. And indeed, they have. Today, that changes thanks to the work of Tobias Weyand, a computer vision specialist at Google, and a couple of pals. These guys have trained a deep-learning machine to work out the location of almost any photo using only the pixels it contains.

116 comments

Min score:

Reason:

Sort:

Can it figure out where goatse was taken? by Anonymous Coward · 2016-02-27 11:33 · Score: 4, Insightful

Can it figure out where the notorious goatse photo was taken?
1. Re: Can it figure out where goatse was taken? by johnsnails · 2016-02-27 11:49 · Score: 3, Funny
  
  Not where bit it can determine it was taken after the London 2012 Olympics as it depics one of the finalist logos. https://encrypted-tbn1.gstatic...
2. Re:Can it figure out where goatse was taken? by Anonymous Coward · 2016-02-27 12:45 · Score: 0
  
  Uranus
3. Re:Can it figure out where goatse was taken? by Applehu+Akbar · 2016-02-27 13:17 · Score: 1
  
  Can it figure out where the notorious goatse photo was taken?
  I thought everyone knew it was Christmas Island.
4. Re:Can it figure out where goatse was taken? by wbr1 · 2016-02-27 13:40 · Score: 1
  
  Goatse as first post. Is this 1998? And yes I was here then under other guise.
  
  --
  Silence is a state of mime.
5. Re:Can it figure out where goatse was taken? by ls671 · 2016-02-27 16:15 · Score: 1
  
  Only if GPS coordinates were embedded in the picture. In short, they are cheating ;-)
  http://osxdaily.com/2015/05/08...
  
  --
  Everything I write is lies, read between the lines.
6. Re:Can it figure out where goatse was taken? by Anonymous Coward · 2016-02-27 17:09 · Score: 0
  
  No. But it can determine what he ate in the last 12 - 24 hrs.
7. Re: Can it figure out where goatse was taken? by Anonymous Coward · 2016-02-27 19:00 · Score: 0
  
  Time paradox detected
Indoor photo location by Anonymous Coward · 2016-02-27 11:35 · Score: 1

I'm well traveled enough to recognize certain kinds of indoor fixtures as being typical for specific countries, or at least regions. One giveaway are electrical sockets. You can learn a lot from the details on those. Window shapes, frames and latches also reveal a lot.
1. Re:Indoor photo location by Anonymous Coward · 2016-02-27 17:17 · Score: 0
  
  Many countries use the same or very similar electrical sockets. Even more so window shapes and frames.
2. Re:Indoor photo location by mikael · 2016-02-27 20:25 · Score: 1
  
  There are some variations. Traditional French buildings tend to have high rectangular windows with frames that open vertically so you can let all the fresh air in. They also have frames for flower boxes at the bottom. Another European feature is to have shutters that can be open against the wall or closed over the window. Modern versions have shutters that roll down over the windows and doors (French legislation for any property that is uninhabited).
  
  --
  Vintage computer adverts: http://www.vintageadbrowser.com/computers-and-software-ads
Almost all = not that many by gweihir · 2016-02-27 11:38 · Score: 2

In mathematics, "almost all" means "all, except for a finite number". As the number of pictures in this universe is finite, this could even mean this "magic" machine cannot do it at all. I guess the headline was written in this spirit.

--
Most ACs are not even worth the keystrokes to insult them. Be generically insulted by this and ignored otherwise.
1. Re:Almost all = not that many by ClickOnThis · 2016-02-27 12:35 · Score: 2
  
  In mathematics, "almost all" means "all, except for a finite number". As the number of pictures in this universe is finite, this could even mean this "magic" machine cannot do it at all. I guess the headline was written in this spirit.
  I'm not sure the author was using the phrase in a mathematical sense. More likely it was in the probabilistic sense of a very high percentage of successes in a number of finite trials.
  BTW, "almost all" in mathematics (specifically, real analysis) can also mean "all, except for a set of zero measure." Consider for example, the Cantor set which has an uncountably infinite number of members lying in the interval [0,1], yet it has measure 0. So, if you construct a function f(x) on [0,1] that has the value 1 at values of x in the Cantor set, but 0 otherwise, then you can say f(x) = 0 at almost all values of x.
  
  --
  If it weren't for deadlines, nothing would be late.
2. Re:Almost all = not that many by Black+Parrot · 2016-02-27 12:36 · Score: 1
  
  In mathematics, "almost all" means "all, except for a finite number".
  Does it always mean that, or almost always mean that?
  
  --
  Sheesh, evil *and* a jerk. -- Jade
3. Re:Almost all = not that many by Dahamma · 2016-02-27 13:02 · Score: 1
  
  I'm not sure the author was using the phrase in a mathematical sense. More likely it was in the probabilistic sense of a very high percentage of successes in a number of finite trials.
  Not really... the author was mostly using it in the Internet clickbait sense.
  If you try it out, it's basically guessing the location of a Google street view image on Google maps. Hardly "determining the location of any image".
  Still, it is a lot of fun. Somehow I scored > 10000 points on the "Paris" map even though I was there my first (and only) time last year. Guess it helps to have walked the shit out of a city...
4. Re:Almost all = not that many by Anonymous Coward · 2016-02-27 13:20 · Score: 0
  
  It almost never means that.
5. Re:Almost all = not that many by luciarowlands · 2016-02-27 14:21 · Score: 1
  
  This is very true, the number of pictures out there is actually very large though. Mid 2014, we humans were posting 1.8 BILLION pictures a day on Internet. I guess it's doubled since then. Most of the data is now created automatically by Internet of Things connected objects (your Fitbit, etc) and I guess the same trend will apply to photography. Makes sense to have automated processes / tools to help us make sense of all this nonsense. The tricky part is, as always with tech, how will they protect our privacy and make sure our personal data extracted from pics are not being sold to insurance companies or other nice entities :)
6. Re:Almost all = not that many by Anonymous Coward · 2016-02-27 16:38 · Score: 0
  
  Yes, (2^24)^(2^32 * 2^32) is technically finite. But it's also more than the number of Hydrogen atoms in the visible universe raised to the 60th power. No man-made computing device could ever enumerate every possible 24-bit RGB x 32-bit width x 32-bit height images before the heat death of the universe. In that sense it's effecively "countably infinite" when it's impossible to count them all before the literal end of time. Besides, nothing's stopping us from making a binary image format that uses a univeral coding like elias delta to allow image width and height of any conceivable size.
7. Re:Almost all = not that many by gweihir · 2016-02-27 18:08 · Score: 1
  
  I am very sure he was not. I was though.
  
  --
  Most ACs are not even worth the keystrokes to insult them. Be generically insulted by this and ignored otherwise.
8. Re:Almost all = not that many by Anonymous Coward · 2016-02-27 18:11 · Score: 0
  
  Correction: "almost all" in mathematics means the same as in English. It's an informal expression that means, "almost all". The term "almost everywhere" is precisely defined, and almost completely unrelated to this comment.
9. Re:Almost all = not that many by squiggleslash · 2016-02-28 01:45 · Score: 2
  
  On Slashdot, "almost all" and "all" are actually synonyms, which is why you'll get flamed if you say something like "Almost all Foobar widgets cost over $10" by people complaining that they found a single Foobar widget for $9.
  
  --
  You are not alone. This is not normal. None of this is normal.
10. Re:Almost all = not that many by K.+S.+Kyosuke · 2016-02-28 10:11 · Score: 1
  
  In mathematics, "almost all" means "all, except for a finite number".
  
  It could easily be an infite number as well, especially if the "all" is uncountable in the first place.
  
  --
  Ezekiel 23:20
11. Re:Almost all = not that many by Anonymous Coward · 2016-02-28 10:15 · Score: 0
  
  In mathematics, "almost all" means "all, except for a finite number". As the number of pictures in this universe is finite, this could even mean this "magic" machine cannot do it at all. I guess the headline was written in this spirit.
  You're a pedant troll. There is a difference between maths and everyday language. Consider the words "group" and "ring"; these words mean completely different things in maths and English.
  Every time something exciting gets announced on Slashdot, some cynical arsehole always has to shit on it, just because.
12. Re:Almost all = not that many by gstoddart · 2016-02-28 14:05 · Score: 1
  
  In mathematics, "almost all" means "all, except for a finite number"
  Hmmm ... did "mostly kinda sorta" become a rigorous mathematical construct when I wasn't looking?
  
  As the number of pictures in this universe is finite
  While waiting for a seat in a restaurant last night, I watched 3 teenage girls taking selfies. I assure you, this is not true.
  And, as a bit of a photography junkie who does a lot of macro photographs and weird things which aren't just straight representational and therefore often don't have much in the way of reference points ... I can say that there will be entire classes of types of photos in which this "magic" machine won't do shit -- like anything with a shallow depth of field in which stuff quickly isn't in focus outside the subject or doesn't provide a lot of other contextual clues.
  And, finally, after years of giving the wife the camera to take pictures, sorting out where a specific picture of a thumb was taken can be almost impossible, even if you have access to the EXIF data. Most photogenic thumb ever. :-P
  
  --
  Lost at C:>. Found at C.
False Title by OpinOnion · 2016-02-27 11:41 · Score: 2

It can't possibly determine the location of any image. Many will simply have too little location data to form a reasonable guess.
1. Re:False Title by Anonymous Coward · 2016-02-27 11:51 · Score: 0
  
  It is in the editor's interest to exaggerate.
  Be that as it may, this is a big step forward for AI. The knowledge work is the final frontier of labor automation. The new world will be amazing.
2. Re: False Title by Anonymous Coward · 2016-02-27 11:51 · Score: 0
  
  Isis are fucked
3. Re:False Title by ClickOnThis · 2016-02-27 12:26 · Score: 1
  
  It can't possibly determine the location of any image. Many will simply have too little location data to form a reasonable guess.
  Precisely. That's why TFS and TFA qualify it by saying it can determine the location of almost any image.
  
  --
  If it weren't for deadlines, nothing would be late.
4. Re: False Title by Anonymous Coward · 2016-02-27 12:40 · Score: 0
  
  Sound stage in Tel Aviv. I suspected as much.
5. Re:False Title by Pseudonym · 2016-02-27 12:50 · Score: 1
  
  On the other hand, it could settle quite a few questions. Give it shots of the moon landings or the alien autopsy.
  
  --
  sub f{($f)=@_;print"$f(q{$f});";}f(q{sub f{($f)=@_;print"$f(q{$f});";}f});
6. Re:False Title by Dahamma · 2016-02-27 13:03 · Score: 0
  
  Which is also totally untrue. It should have said "it can determine the location of almost any image ON GOOGLE STREET VIEW".
7. Re: False Title by guruevi · 2016-02-27 13:14 · Score: 0
  
  But that's not really a neural network then. To "find" subsets of an image within a set of other images, even a (large) database of them is relatively trivial (from a coding perspective) even if things like color, lighting and compression change and doesn't require any "special sauce" like AI or neural networks.
  What's next: Apple Siri can guess songs it hears from their iTunes catalog? FaceBook can identify faces within pictures? YouTube has a search function.
  
  --
  Custom electronics and digital signage for your business: www.evcircuits.com
8. Re:False Title by Anonymous Coward · 2016-02-27 13:19 · Score: 0
  
  It is in the editor's interest to exaggerate.
  Be that as it may, this is a big step forward for AI. The knowledge work is the final frontier of labor automation. The new world will be amazing.
  Your new world will be filled with billions of humans unable to find a job.
  And our societal structure will not move fast enough to accept a human working a 10-hour workweek and getting paid for 40. Greed will take over, as if the continued separation between the 99% and the 1% isn't evidence enough.
9. Re: False Title by Anonymous Coward · 2016-02-27 13:26 · Score: 0
  
  Lol
10. Re: False Title by Dahamma · 2016-02-27 16:14 · Score: 1
  
  But that's not really a neural network then
  Wha? It's either using a neural network or it isn't. Whether it's a neural network or not doing the computing has nothing to do with the results or the efficiency of the search vs more traditional computing methods.
  And I wasn't commenting on the methods or even the effectiveness of the program, just the (probably clueless non-tech) editor's assertion that it applied to "almost any image".
11. Re: False Title by ClickOnThis · 2016-02-27 16:20 · Score: 1
  
  But that's not really a neural network then
  Wha? It's either using a neural network or it isn't. Whether it's a neural network or not doing the computing has nothing to do with the results or the efficiency of the search vs more traditional computing methods.
  If the neural network hasn't been trained in some way on the images it sees, then it's not fair to ask it to recognize them.
  The same criterion applies to a human being.
  
  --
  If it weren't for deadlines, nothing would be late.
12. Re:False Title by Mattcelt · 2016-02-27 17:20 · Score: 1
  
  I can tell with 99.9999999999% accuracy where ANY photo was taken, without even looking at the pixels:
  Earth.
  What the article fails to mention is any reasonable measure of geographical range. It's one thing to say, 'this photo was taken in China'. It's quite another to say, 'this photo of cat food was taken in Bob Smith's kitchen at 22 Kings Mews, London, SW1'.
  The real news here is the claim that the software/network can pinpoint as well as a human can. And I'd like to see that tested.
13. Re:False Title by angel'o'sphere · 2016-02-28 02:15 · Score: 1
  
  Considering that plenty of pictures will be "Golden Gate" and "Eiffel Tower" I would assume you would be surprised how many places you actually can guess.
  Don't have to been there, movies are enough.
  
  --
  Cost free eBook I read (by iBook/Kobo/Amazon/ObookO/Gutenberg etc.): "The Green Odyssey" by Philip Jose Farmer.
14. Re: False Title by angel'o'sphere · 2016-02-28 02:33 · Score: 1
  
  To "find" subsets of an image within a set of other images, even a (large) database of them is relatively trivial (from a coding perspective) even if things like color[...]
  Then sketch us your algorithm here and farm in your Nobel Prize.
  
  --
  Cost free eBook I read (by iBook/Kobo/Amazon/ObookO/Gutenberg etc.): "The Green Odyssey" by Philip Jose Farmer.
15. Re:False Title by Mike610544 · 2016-02-28 07:42 · Score: 1
  
  What the article fails to mention is any reasonable measure of geographical range.
  "PlaNet is able to localize 3.6 percent of the images at street-level accuracy and 10.1 percent at city-level accuracy," say Weyand and co. Whatâ(TM)s more, the machine determines the country of origin in a further 28.4 percent of the photos and the continent in 48.0 percent of them.
  
  The real news here is the claim that the software/network can pinpoint as well as a human can. And I'd like to see that tested.
  "In total, PlaNet won 28 of the 50 rounds with a median localization error of 1131.7 km, while the median human localization error was 2320.75 km."
  
  I know it's standard practice here to comment without reading the article. But why specifically stare that it doesn't contain exactly the information that it does?
  
  --
  ... also, I can kill you with my brain.
16. Re:False Title by Mattcelt · 2016-02-28 07:49 · Score: 1
  
  Em... right, then. You are correct, I did not RTFA.
  I meant to write "post", not "article".
  But that does not excuse my laziness or misspeaking. My apologies.
17. Re: False Title by guruevi · 2016-02-28 08:59 · Score: 1
  
  How the hell do you think face (or any sort of image) recognition works? Never heard of license plate scanners? Those things are relatively trivial.
  
  --
  Custom electronics and digital signage for your business: www.evcircuits.com
18. Re:False Title by Anonymous Coward · 2016-02-28 14:06 · Score: 0
  
  I think I located one picture because the wall matched the wall in one building I happened by, so at least I can search for architect, quarry, etc. Nope, I cannot simply infer it is the same city, architect may have imported from a quarry then went to other city to do a same style building.... I do not think these guys have ALL that information, but MAYBE this particular case is indeed incorporated in their knowledge bases. So at best they may say USA and still not know where exactly the picture was taken.
19. Re: False Title by angel'o'sphere · 2016-02-29 01:18 · Score: 1
  
  And? You seem not to know the difference in trivia between a license plate scanner and your "simple proposal" of simple picture matching ;D
  How the hell do you think face (or any sort of image) recognition works? 99% of it works by "knowing what you are looking for and figuring where it is in the picture" ...
  Those things are relatively trivial. No, they are not. There are dozens or 100ds of algorithms involved to look for a single feature. I doubt you know any of them.
  
  --
  Cost free eBook I read (by iBook/Kobo/Amazon/ObookO/Gutenberg etc.): "The Green Odyssey" by Philip Jose Farmer.
Try it yourself by mobby_6kl · 2016-02-27 11:41 · Score: 4, Interesting

Well the "Guess the location" thing, not the NN :)
There's a site that basically opens StreetView at random around the world and asks you to place it on the world map. As the summary explains, you can use a number of clues to generally place photos surprisingly accurately. Used to play this occasionally at work, we really liked that it challenged you to think about all these things that you know about the countries and regions around the world.
https://geoguessr.com/
1. Re:Try it yourself by Anonymous Coward · 2016-02-27 12:30 · Score: 0
  
  I hate you. I have spent the last 5 hrs in Namibia because of that link. It was fun, but I finally gave up and found I was 500 miles away from where I thought I was.
  This is going to turn into an addiction real quick.
2. Re:Try it yourself by Anonymous Coward · 2016-02-27 13:06 · Score: 0
  
  by mobby_6kl (668092) on Sunday February 28, 2016 @12:41AM
  by Anonymous Coward on Sunday February 28, 2016 @01:30AM
  
  the last 5 hrs
  Yeah.. sure.
3. Re:Try it yourself by JazzLad · 2016-02-27 14:21 · Score: 1
  
  Doctor, is that you?
  
  --
  "If you have nothing to hide, you have nothing to fear." - Every fascist, ever
4. Re:Try it yourself by adolf · 2016-02-27 17:02 · Score: 1
  
  As the summary explains, you can use a number of clues to generally place photos surprisingly accurately.
  Feedlot? John Deer tractor mowing? White lines instead of yellow? A squished rabbit on the road?
  (I wasn't sure until I saw the rabbit.)
  
  --
  Kid-proof tablet..
5. Re:Try it yourself by mobby_6kl · 2016-02-27 21:11 · Score: 1
  
  Feedlot? John Deer tractor mowing? White lines instead of yellow? A squished rabbit on the road?
  (I wasn't sure until I saw the rabbit.)
  If you're the one who hit that rabbit you could probably place the image pretty damn accurately!
  Generally, if it looks like America but with fewer trucks and everyone drives on the opposite side, it's Australia. If looks quite like America but something's slightly off, that's Canada. If everyone drives on the opposite side and it's cloudy, wet, and depressing, thats's the UK. If it just looks depressing, that's Russia.
  Here's my result from the fist try after not playing for a while. My biggest challenge seems to be the inability to differentiate between Australian states.
  Also I just now finally checked out the actual article and this game is actually mentioned there. Just shows that RTFA is for suckers.
6. Re:Try it yourself by MrL0G1C · 2016-02-27 21:46 · Score: 1
  
  I just hunted around for street signs and googled them! Cheating no doubt but fiendishly difficult otherwise when you've got nothing to go on other than a road and some fir trees.
  
  --
  Waterfox - a Firefox fork with legacy extension support, security updates and better privacy by default.
7. Re:Try it yourself by Anonymous Coward · 2016-02-27 23:59 · Score: 0
  
  Yes, now take your meds, and transfer that money to my account like we talked about.
8. Re:Try it yourself by mobby_6kl · 2016-02-28 00:04 · Score: 1
  
  That's not much fun though, is it :)
  If you have realistic expectations, often you don't really need much more than a road and some trees - you can tell the country by the road signs (like, speed limits and stuff, not city names :D) and lane markings and the region by the trees. But sometimes you're unlucky and get a dirt road in the middle of a pine forest or something, then you're screwed.
  One place I would expect the AI to do a better job is identifying cities. I can ID NYC, DC, Paris or Berlin, but I can't tell one anonymous shithole from another, country & region-level is probably the best I could hope for. The AI , on the other hand, would've probably seen that very city before.
  I screwed up my link in the other post above, here's my last result: http://imgur.com/TqTkDYE
9. Re:Try it yourself by MrL0G1C · 2016-02-28 01:31 · Score: 1
  
  Actually it's still a challenge, I had to type in Russian letters, zoom in on a tv screen, hunt for road signs, still make guesses based on countryside, road markings etc. My places were more diverse, I got Argentina middle of nowhere, ditto for Iceland, Russia and Canada.
  
  --
  Waterfox - a Firefox fork with legacy extension support, security updates and better privacy by default.
10. Re:Try it yourself by nickersonm · 2016-02-28 03:06 · Score: 1
  
  Per the original paper, that's what they use to train the NN and as a goodness metric.
11. Re:Try it yourself by adolf · 2016-02-28 06:00 · Score: 1
  
  No vehicles in sight, except for the mowing tractor. No road signs nearby, either.
  It was Australia. Rabbits are ridiculously common there.
  It's an odd game, it is habit-forming, and it gets easier as it gets played... It should be part of a high school geography or social studies class.
  And there goes my afternoon.
  
  --
  Kid-proof tablet..
12. Re:Try it yourself by Hentes · 2016-02-28 08:25 · Score: 1
  
  There are many kinds of random. The site chooses a photo without taking into account where it was taken, and because the density of photos varies greatly with location, even a simple statistical algorithm using no data from the images could get fairly accurate at guessing.
13. Re:Try it yourself by sysrammer · 2016-02-28 19:31 · Score: 1
  
  Thank you. I had run out of things to waste time on the internet.
  
  --
  His ignorance covered the whole earth like a blanket, and there was hardly a hole in it anywhere. - Mark Twain
Almost any photo = 10.1% by ebcdic · 2016-02-27 11:46 · Score: 4, Informative

According to the article, it can identify 10.1% of the Flickr images it was tested on at "city-level" accuracy.
1. Re:Almost any photo = 10.1% by Anonymous Coward · 2016-02-27 11:48 · Score: 0
  
  Whew, my exhibitionist dick pics are safe for now!
2. Re:Almost any photo = 10.1% by Anonymous Coward · 2016-02-27 13:00 · Score: 1
  
  NOBODY reads the articles. The brilliant minds on Y Combinator news also only read the headline.
3. Re:Almost any photo = 10.1% by thegarbz · 2016-02-27 18:34 · Score: 2
  
  Compared to a human that's pretty dam good.
4. Re: Almost any photo = 10.1% by Anonymous Coward · 2016-02-27 19:34 · Score: 0
  
  Speak for yourself. I knew my vouyerist ways and photograph would cause me trouble but never like this!
5. Re:Almost any photo = 10.1% by Anonymous Coward · 2016-02-28 01:48 · Score: 0
  
  this is why street view, picasa, and google images exist.. so google can mine the images and metadata.. same for yahoo and its image search and flickr, same for facebook, its photos and instagram.
6. Re:Almost any photo = 10.1% by GerbilKor · 2016-02-28 03:35 · Score: 1
  
  Speak for yourself. All of my exhibitionist dick pics have the signs for the nearest cross-road in the background. How could I have known!?
7. Re:Almost any photo = 10.1% by AthanasiusKircher · 2016-02-28 07:51 · Score: 1
  
  Compared to a human that's pretty dam good.
  Depends on what you're trying to measure.
  Computers are always faster and better at dealing with large quantities of data. For example, if I want to search a book-length text, my computer can instantly find all strings matching a given pattern. Obviously, a human could do this too, but it would take a while, and if rushed, the human might miss some.
  But I don't think we'd say that a simple search for a string indicates "intelligence" in the AI sense, no matter how much faster or more accurate a computer might be at doing it.
  Undoubtedly you're correct that humans probably wouldn't be able to identify even 10% of random photos in random cities. But the question raised by the findings here is whether the "AI" has achieved that statistic just because it's doing something rather trivial and just has the ability to work more quickly and with a larger database (like the text-search analogy above).
  Or, is the "AI" actually doing something more "intelligent" and sophisticated?
  To answer that latter question, we'd be better off trying a comparison with a human who had lived in 3 or 4 cities over a span of a few recent years. (Better yet, choose a human who is likely to have traveled a lot around that city and probably is quite familiar with surroundings in various parts, like a taxi driver or something.) Will the computer algorithm here perform better than the human in selecting the correct cities for photos under those conditions?
  If so, then we can get closer to saying there's a rather sophisticated "AI" here that might be "better than human" in this general pattern recognition task, rather than just a computer with access to more data than a human can hold in memory at once.
geoguessr by Anonymous Coward · 2016-02-27 11:55 · Score: 0

Addicted to that game.
Aw gee golly swell! by Anonymous Coward · 2016-02-27 12:03 · Score: 0

Gee, he and a couple of pals! How swell is that! The American dream!
Except that this was done decades ago by the military, and these guys are simply riding their privilege to success on the coattails of the MIC to which they have sucked up. From the largest intelligence entity in the known universe.
Uh oh by rampant+mac · 2016-02-27 12:13 · Score: 2

Guess us old-timers will finally find out where the goatse guy is hiding out. :(

--
I like big butts and I cannot lie.
1. Re:Uh oh by Anonymous Coward · 2016-02-27 12:19 · Score: 0
  
  You're looking for the goatse guy?
  You are a mac user...good luck, tie a rope around your waist and to the bedpost.
2. Re:Uh oh by OzPeter · 2016-02-27 13:15 · Score: 1
  
  Guess us old-timers will finally find out where the goatse guy is hiding out. :(
  He's probably sharing a house with tub girl
  
  --
  I am Slashdot. Are you Slashdot as well?
3. Re:Uh oh by Frosty+Piss · 2016-02-27 13:43 · Score: 1
  
  Guess us old-timers will finally find out where the goatse guy is hiding out. :(
  He's probably sharing a house with tub girl
  Sorry, but it's the Lemon Party dude. Now go wash your eyes out with bleach.
  
  --
  If you want news from today, you have to come back tomorrow.
4. Re:Uh oh by Anonymous Coward · 2016-02-27 14:08 · Score: 0
  
  Guess us old-timers will finally find out where the goatse guy is hiding out. :(
  He goes under the name kirkj. Just ask him.
5. Re:Uh oh by Bite+The+Pillow · 2016-02-27 20:39 · Score: 1
  
  Duh, goatse guy doesn't have a mailing address, he has a PO box. In his ass.
  You can mail him, but you can't female him.
  He's not hiding out, his ass is everywhere.
  Last time someone said, "Up yours," he said, "You'll need at least two."
  And if A.I. ever needs to be put down, and a paradox fails, "Where is goatse guy" should give you a few days to escape while a computer tries to figure out "Wtf is this," "I can't even," and finally, "How many jelly bellies worth of excavated arsehole is that, approximately?"
6. Re:Uh oh by sysrammer · 2016-02-28 11:12 · Score: 1
  
  ...in a maze of twisty little passages, all alike.
  
  --
  His ignorance covered the whole earth like a blanket, and there was hardly a hole in it anywhere. - Mark Twain
Take a shot if.... by Anonymous Coward · 2016-02-27 12:26 · Score: 1

you watch a movie with it and it returns Vancouver
Great! by Black+Parrot · 2016-02-27 12:30 · Score: 2

Now, where does it say the moon landings were filmed?

--
Sheesh, evil *and* a jerk. -- Jade
1. Re:Great! by Anonymous Coward · 2016-02-27 13:08 · Score: 0
  
  Now, where does it say the moon landings were filmed?
  The moon.
  Next question.
2. Re:Great! by Anonymous Coward · 2016-02-27 13:22 · Score: 0
  
  Now, where does it say the moon landings were filmed?
  We have the capability today to point telescopes at the moon and see the lunar tracks from previous missions, and we're still going to dredge this conspiracy theory shit up?
  I can only hope this was a poor attempt at humor.
3. Re:Great! by Anonymous Coward · 2016-02-27 16:01 · Score: 0
  
  We have the capability today to point telescopes at the moon and see the lunar tracks from previous missions, and we're still going to dredge this conspiracy theory shit up?
  No we don't.
  We do have satellites in orbit around the moon that can image the tracks, and the LEM landing stage, and science packages, though.
4. Re:Great! by penguinoid · 2016-02-27 17:07 · Score: 1
  
  Now, where does it say the moon landings were filmed?
  The moon.
  Do you have any evidence of that? It wouldn't surprise me if that AI doesn't even know the moon exists.
  
  --
  Don't waste your vote! Vote for whoever you want, unless you live in a swing state it won't matter anyways
Re: Can the neural network also help with... by mrmatthewcarlson · 2016-02-27 12:40 · Score: 1

You're not worth a dollar an hour.
Picture of ... by PPH · 2016-02-27 12:50 · Score: 1

... my car keys.

--
Have gnu, will travel.
1. Re:Picture of ... by TeknoHog · 2016-02-27 13:14 · Score: 1
  
  ... my car keys.
  They're in your car, and it's moving at exactly 88 mph. Not even Google can beat physics.
  
  --
  Escher was the first MC and Giger invented the HR department.
Bite my shiny metal ass by easyTree · 2016-02-27 12:51 · Score: 1

How many of the human contestants had been trained using 91 million tagged images specifically for this purpose?

--
Requiem for the American Dream
Google NSA by Anonymous Coward · 2016-02-27 12:58 · Score: 0

'Nuff said.
It's Weyland by Anonymous Coward · 2016-02-27 13:06 · Score: 0

Pretty sure that's a typo.
Find my arse by Anonymous Coward · 2016-02-27 13:24 · Score: 0

Submit a picture of your bhole and see if it can find where your arse is located
How clever? by manu0601 · 2016-02-27 14:06 · Score: 1

How will it cope with an indoor picture from Australia with an Eiffel tower picture hanging on the wall?
Any human will conclude the Eiffel tower here does not mean the picture is from Paris, but a computer?
1. Re:How clever? by Anonymous Coward · 2016-02-27 14:55 · Score: 0
  
  Why don't you try it and report back to us?
Artists can sense location by the quality of light by dsmatthews9379 · 2016-02-27 15:27 · Score: 4, Interesting

Perhaps not all of them but it is a thing many talk about, that different locations have a distinct quality to the natural light there. Not just location either, but seasonal variation too. It would not surprise me if the NN was also sensitive to these clues. Which makes me wonder can Weyand el al get their NN to tell them additional things about the image such as time of day and day of year, perhaps even in some cases actual year because it can cross reference images in the same location with known dates against an image that it has geo-located and know what the weather (lighting conditions) were for that point in time near that location.
Organic by jargonburn · 2016-02-27 15:44 · Score: 1

This is why I only use organic pixels for my photos. None of those unnatural digital markers contaminating my files!
Doesn't Simulate Human Reasoning by Anonymous Coward · 2016-02-27 15:46 · Score: 0

Yeah, but in the end, the neural net still has no explicit "kinds of knowledge about the world such as the type and language of signs on display, the types of vegetation, architectural styles, the direction of traffic, and so on." It just kept guessing with brute force, got feedback, and heuristically mapped results to visual patterns. These colored regions in the images with a striking similarity to each other have a tendency to be associated with the location of Russia. It doesn't what signs are. What letters are. Much less the Russian alphabet, its words, syntax, or grammar, or why certain words are chosen to be on signs vs. others. This is not at all similar to how a human becomes good at gleaning locations, too. The computer is reasoning completely bottom-up, whereas a human would mix top-down and bottom-up.
Re: Can the neural network also help with... by Anonymous Coward · 2016-02-27 17:30 · Score: 0

So he's worth less than his mom?
Criminal tracking by phorm · 2016-02-27 18:32 · Score: 1

I wonder if they're planning on using this to track criminals such as child pornography creators etc online. It would certainly be a good use for this. I doubt it's accurate enough to capture a house etc (unless somebody was dumb enough to have other pics in the same house/room), but it might help regionalize it.
Re: Can the neural network also help with... by neo8750 · 2016-02-27 19:36 · Score: 1

You saying she raised her prices?
Re: Can the neural network also help with... by Anonymous Coward · 2016-02-27 20:53 · Score: 0

You created an account just so you could post that? Wow! What a loser!
Clearly you're not worth a dollar per hour. Not even close.
False. by eWarz · 2016-02-27 22:10 · Score: 1

Google claims to have this so-called 'deep-learning machine' can identify the location of any photo? I will bet Google 1 BILLION dollars that they can't identify the locations of most of the photos I have...all pre 90s photos taken all over the US with absolutely NO landmarks in the background or other key deciding factors. Someone at Google got waaay too cocky about their 'slightly improved' algorithm. On the other hand, I once knew a guy who invented a machine that could generate 400% of the power required to power itself...
"machines" would struggle? by Anonymous Coward · 2016-02-28 02:08 · Score: 0

"Humans spend a lifetime picking up these kinds of geolocation cues. So it's easy to think that machines would struggle with this task. And indeed, they have."
Dude, humans do all of the struggling, don't put the blame on the machines... Had they been conceptualized & implemented correctly so they could do the task, so would they.
Slashdot clickbait headline by Jiro · 2016-02-28 03:49 · Score: 1

The results make for interesting reading. To measure the accuracy of their machine, they fed it 2.3 million geotagged images from Flickr to see whether it could correctly determine their location. "PlaNet is able to localize 3.6 percent of the images at street-level accuracy and 10.1 percent at city-level accuracy," say Weyand and co. Whatâ(TM)s more, the machine determines the country of origin in a further 28.4 percent of the photos and the continent in 48.0 percent of them.
Oh,. boy, 3.6 percent. Yeah, any image, not.
(Also, 48% continent is worthless because you could get some percentage just by guessing USA or Europe each time.)

In total, PlaNet won 28 of the 50 rounds with a median localization error of 1131.7 km, while the median human localization error was 2320.75 km
This is worthless too. Not just because 1000 instead of 2000 is still pretty big, but because you could improve performance over humans just by doing things like "if you only know the country, guess the population center", which most humans won't do.
EOP by fyngyrz · 2016-02-28 08:11 · Score: 1

Many types of change are stressful, dangerous, etc. That doesn't mean a change shouldn't happen. Sometimes, as I assert is the case with menial labor, the status quo has no particular merit. The sooner we get to an economy of plenty, the better off humanity will be. Yes, the 1% will likely do their very best to stand in the way, but it is my hope that change simply steamrollers them.

--
I've fallen off your lawn, and I can't get up.
Free help by fyngyrz · 2016-02-28 08:15 · Score: 1

I can't tell one anonymous shithole from another
You're on slashdot.
No, no need to thank me. I'm into public service.

--
I've fallen off your lawn, and I can't get up.
so can I by Anonymous Coward · 2016-02-28 13:22 · Score: 0

Give me a few billion images taken in random locations around the world, and I could probably score within the 99th percentile of accuracy by saying they were taken on the planet earth!. Very few images are available from outside the atmosphere, so they would bare be a blip in the error stats
Challenge accepted... by evilviper · 2016-02-28 14:28 · Score: 1

Okay Google... Tell me exactly where on the planet THIS PHOTO was taken.
I'll wait...

--
Slashdot gets worse every day... Pipedot: News for nerds, without the corporate slant
Great news for 4chan by o.sironi · 2016-02-28 18:26 · Score: 1

Presto doxed!
Track down paedophiles! by Anonymous Coward · 2016-02-29 08:37 · Score: 0

A massive chance to save so many of the victims of paedophiles, and return trafficked children to their real homes, will law enforcement jump at this opportunity god I hope so. There are few things that can offer real hope to these children, and by the sounds of this it would be a massive leap in the right direction.
Re: Can the neural network also help with... by Anonymous Coward · 2016-03-04 14:18 · Score: 0

Fuck you.
Re: Can the neural network also help with... by Anonymous Coward · 2016-03-04 14:34 · Score: 0

You contribute nothing of value to humanity.
Re: Can the neural network also help with... by Anonymous Coward · 2016-03-13 06:47 · Score: 0

You're an idiot. Go to hell.
Re: Can the neural network also help with... by Anonymous Coward · 2016-03-13 06:55 · Score: 0

I fucked your mother last night and the bitch screamed. She wasn't a good fuck, though, not even worth a few pennies.
Re: Can the neural network also help with... by Anonymous Coward · 2016-03-13 11:09 · Score: 0

Fuck off! Loser!
Re: Can the neural network also help with... by Anonymous Coward · 2016-03-13 12:57 · Score: 0

I tried to have sex with your mom but the bitch is too fat and there's too much flab in the way.
Re: Can the neural network also help with... by Anonymous Coward · 2016-03-13 23:37 · Score: 0

You're an idiot. GTFO.