Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Google's Street View computer vision can beat reCAPTCHA with 99% accuracy

googleonlinesecurity.blogspot.com

Post reply on HN
Print view View original

71–80 of 159 posts

Re: Google's Street View computer vision can beat reCAPTCHA with 99% accuracy

#71

Post by bentcorner »

zwegner wrote: post #45

This particular issue (AI performance on captchas) is really quite fascinating. It's an arms race, but the problem is, only one side can win. Google is claiming they have improved their system in some (understandably) unspecified way, but there's only so far this can go. Captchas need to detect whether someone is human, but it has to work for everyone, ideally, even those with disabilities. Any simple task a human ca…

Ask for x BTC on signup, on the promise that it is returned in y weeks.

I was going to propose that it's only returned on good behavior, but that gives sites an incentives to harshly police their new user population.

Ideally you'd have some sort of service that both parties trust that is holding the BTC.

Re: Google's Street View computer vision can beat reCAPTCHA with 99% accuracy

#73

Post by zwegner »

Lambdanaut wrote: post #67
zwegner wrote: post #65

Earlier quoted context omitted.

Presumably a service would only allow one (or a few) accounts per actual person. The end user will still need some amount of security. I had in mind some challenge-response mechanism, like: Person signs government ID number with public key, presents it to authority Authority signs public key and asserts it belongs to a real human Service presents random number at signup time Person signs the random number with the sa…

I'm not entirely sure how that would mitigate people from passing the government signed public key to their bot and letting it take over from there.

That's what I was referring to with services only allowing one (or N) accounts per person. A spammer could use the signature multiple times, but would need new email providers for every single account.

And as an uncle (?) comment points out, there could possibly be blacklists of signatures of spammers shared between providers. That definitely would cause some new problems, of course, but who knows what the future may hold...

Re: Google's Street View computer vision can beat reCAPTCHA with 99% accuracy

#74

Post by ralfn »

Keyframe wrote: post #54
return0 wrote: post #42

Earlier quoted context omitted.

The endgame solution is to let humans judge other humans. In 5 years, you will be interrogated every time you post to HN. The machines are here to make our lives hell...

In 5 years, you will be interrogated every time you post to HN But that's how it already is! :) Eventually spam content will either blend in to be relevant and attractive (we can see hints of it in new youtube ads system, but that's matching to ads only, not generating) or it will be analyzed and scrapped (along with other comments) by other equally intelligent systems. So machines will judge both machines and humans…

When it becomes impossible to distinguish computer intelligence from human intelligence, do we really want to stop the computers from commenting?

If a computer algorithm can create a more informed relevant comment than a person, maybe i would prefer to see the artificial comment.

Re: Google's Street View computer vision can beat reCAPTCHA with 99% accuracy

#75

Post by PeterisP »

Lambdanaut wrote: post #60
zwegner wrote: post #45

This particular issue (AI performance on captchas) is really quite fascinating. It's an arms race, but the problem is, only one side can win. Google is claiming they have improved their system in some (understandably) unspecified way, but there's only so far this can go. Captchas need to detect whether someone is human, but it has to work for everyone, ideally, even those with disabilities. Any simple task a human ca…

The suggestion to "go to a physical location to get a trusted government signature" has deeper flaws as well. What's to stop a bot writer from getting a signature and plugging it into his bot?

1. Unlike e-mail accounts or such, a botwriter wouldn't be able to get thousands of such signatures to go on as soon as the first one is identified and blocked; if he puts his own key in the bot - bam, you (and everyone else you know) can exclude him from your services permanently, and he wouldn't be able to come back with a fresh account unlinked to that one, ever.

2. The signature is linked to a specific person, so if the private key is stolen, then you (a) revoke it; and (b) punish the thief - because obtaining that signature most likely be treated as a felony, unlike spamming comments or other uses for breaking captcha's.

Re: Google's Street View computer vision can beat reCAPTCHA with 99% accuracy

#76

Post by daenz »

zwegner wrote: post #45

This particular issue (AI performance on captchas) is really quite fascinating. It's an arms race, but the problem is, only one side can win. Google is claiming they have improved their system in some (understandably) unspecified way, but there's only so far this can go. Captchas need to detect whether someone is human, but it has to work for everyone, ideally, even those with disabilities. Any simple task a human ca…

> It's an arms race, but the problem is, only one side can win.

I don't disagree, but although this specific problem can be automated, the arms race will continue as newer problems (requiring uniquely human intelligence) are presented. We can keep moving the bar until the singularity.

For example, a captcha: "Pick the most unsafe environment", followed by pictures of a house, a park, a volcano, a bed. To solve this, AI would have to have a mastery of language, a mastery of object identification, and a mastery of the metadata of those objects as they apply to human safety.

Re: Google's Street View computer vision can beat reCAPTCHA with 99% accuracy

#77

Post by id »

akgerber wrote: post #70
ashmud wrote: post #53

Earlier quoted context omitted.

Do people ever intentionally type in nonsense for the non-captcha part, out of spite, for instance? If you could persuade enough people to provide non-answers for the non-captcha part, would Google stop using it (the non-captcha part)? My guess is it's like spam. As long as it works a small fraction of the time, it will still exist.

I did when Recaptcha came out and it was consistently obvious which was the unknown part. It's no longer so obvious, in my experience. I believe ReCaptcha presents the same unknown value to several users, and rejects outliers.

In my experience, it's still extremely obvious. It's always the word that's less readable (and doesn't look like it came from a book). And if you have numbers, it's even easier.

Recently they added special captchas for untrusted users. If they think you might be a bot, you'll get much harder captchas.

http://googleonlinesecurity.blogspot.com/2013/10/recaptcha-j...

Re: Google's Street View computer vision can beat reCAPTCHA with 99% accuracy

#78

Post by stuaxo »

I'm sorry, as a human I have had to fill these street view style captchas in all the time for google, so this is hardly a completely artificial intelligence, humans have done it many many times, in fact I'm sure some of the pictures in the articles have come up.

Re: Google's Street View computer vision can beat reCAPTCHA with 99% accuracy

#79

Post by sukuriant »

daenz wrote: post #76
zwegner wrote: post #45

This particular issue (AI performance on captchas) is really quite fascinating. It's an arms race, but the problem is, only one side can win. Google is claiming they have improved their system in some (understandably) unspecified way, but there's only so far this can go. Captchas need to detect whether someone is human, but it has to work for everyone, ideally, even those with disabilities. Any simple task a human ca…

> It's an arms race, but the problem is, only one side can win. I don't disagree, but although this specific problem can be automated, the arms race will continue as newer problems (requiring uniquely human intelligence) are presented. We can keep moving the bar until the singularity. For example, a captcha: "Pick the most unsafe environment", followed by pictures of a house, a park, a volcano, a bed. To solve this,…

At the same time, though, a human would have to have familiarity with all of those objects in the context of their environment. A house doesn't look the same to everyone, nor does a park, bed, etc. People's life experiences can also affect their responses.

The problem we would then experience is unintentional prejudice and Western-mindedness in our "are you human" queries, similar to a problem that SATs and other standardized testing methods have been criticized for.

Post reply on HN