10
$\begingroup$

Are there any modern techniques of generating textual CAPTCHA (so person needs to type the right text) challenges which can easily fool AI with some visual obfuscation methods, but at the same time human can solve them without any struggle?

For example I'm talking about plain ability of recognising text embedded into image (without considering any external plugins like flash or java, image classification, etc.) and re-typing the text that has been written or something similar.

I guess adding noise, gradient, rotating letters or changing colours are not reliable methods any more, since they can be quickly broken.

Any suggestions or research has been done?

$\endgroup$
  • 4
    $\begingroup$ Isn't this sort of backwards? Usually someone first makes a Captcha that they think can fool a bot, THEN other people start working on solving it automatically. Anything new you might think of will become obsolete very quickly. $\endgroup$ – Disenchanted Lurker Aug 5 '16 at 8:17
  • $\begingroup$ Try to think about what humans can do better than computers. We can reason and we (native speakers) will know nearly every idiom. See this for further ideas. I think that as far as text goes, the best would be to have the user enter text that is analyzed with NLP to match a declared sentiment or perhaps express something. Computers are not very good at drafting clear, well-structured sentences (but I guess most humans aren't great at that either). $\endgroup$ – JakeD Oct 9 '16 at 2:41
  • $\begingroup$ Usually textual CAPTCHA is used to mean that the CATPCHA is presented as text, not that the required user input must be text. For example TextCaptcha. $\endgroup$ – Theraot Oct 10 '16 at 20:30
5
+50
$\begingroup$

It's an interesting question about what makes humans unique. There is a good book on the subject titled What Computers Cant Do by Hubert Dreyfus.

One task that a computer can't handle (for now at least) is ranking important things. For example, CAPTCHA asks you to order a random list of things (small one, five or six items) by importance. This particular exercise requires AI to take decisions (not always rational) based on human judgement.

$\endgroup$
5
$\begingroup$

A method that could possibly work is utilising optical illusions such as one where two lines down a hallway are identical but one seems longer to the human eye, then they could be prompted with a multiple choice question as to the state of the line, which to our eyes looks longer, but to a computer, is still the same length of line. Of course, there is always the issue of people with eye based disabilities not being able to complete them, but different illusions could be used to accommodate that.

Example

$\endgroup$
2
$\begingroup$

Have the user label highlighted objects in video that a state of the art classifier cannot solve

Create a state of the art video classifier. Might as well train it on Google's YouTube-8M video training data. But you will want to continually feed it original video as well.

Have the classifier label as many objects as it can. Have it isolate which objects it can recognize as objects but which it is unable to label.

Have it output videos that outlines the objects. Preferably GIFs, which can be easily embedded in forms.

For 100 of these, ask 100 users what the object is. If 90% of the users agree on the name of an object, add that video to the captcha-set. Call this the pre-trained set.

Every time a user needs to authenticate, show them one of the highlighted objects in a video not from the pre-trained set. If the image has less than 100 showings, record the label and give the user another one from the pre-trained set. If they get it right, let them through, if not, give them another from the pretrained set.

Once the non-pre-trained video has more than 100 showings and more than 90% of the captcha-users agree, add that video to the post-trained set.

Over time, slowly remove the pre-trained set. Put expirations on each video in the post-trained set and remove them after expiration, so that they don't get used too many times.

Ideally, this process would constantly improve the video classifier, keeping it state of the art and slightly ahead of other classifiers. Perhaps it could also favor less common words and objects and more esoteric things, so as to specialize this classifier against other classifiers.

The same could be done for image labeling, but the utility of the video classifier will probably last longer, given advances in AI.

Strictly speaking, though, short of some quantum trickery, there is no captcha system that will not one day be solved by external AI systems.

(edit: oh, I just noticed you specifically said "textual captcha." If that's what you mean, then no I don't think text classification has much mystery left in it. Computers can probably glean text from pictures better than humans now. But techically, the input in the above described captcha system is textual.)

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.