Technologies for Good and Bad

Just as new machine learning techniques can be used for the good, they can also be used for bad purposes.  The below research is been being reported as an example, ways of cracking encryption is another. ...

AI Model Fundamentally Cracks CAPTCHAs
National Public Radio   By Merrit Kennedy

Researchers at the artificial intelligence company Vicarious say they have developed a computer model that can solve Completely Automated Public Turing Tests to Tell Computers and Humans Apart (CAPTCHAs). Vicarious co-founder Dileep George says the model cracks a CAPTCHA's defenses by parsing the text the test presents more effectively than earlier deep-learning models and with less training. George cites the model's use of a Recursive Cortical Network, which is better able to reason about what it is seeing with less training than previous models. He says the network's training entails constructing internal models of the letters it is exposed to, and when a new image is presented it attempts to explain it, trying to define all of the pixels of that new image in terms of the characters it has previously seen. George notes the long-term objective of such experiments "is to build intelligence that works like the human brain. CAPTCHAs were just a natural test for us."

When Computers Become Smarter than Humans

Knowledge@Wharton looks at the often posed question.  The answer is, they already are.  First they did math better, then searched better, connected to people better, played chess better, stored data better  ...     But these are all relatively narrow areas. Becoming less narrow every day.  At what time do all of these areas combine into everything we do?    Even in these still relatively narrow areas. And is that then assumed to be 'smart'?  Smart,  even for humans,  is always in some context.

What if Computers Become Smarter Than Humans?  Nov 22, 2016

You’re ordering tickets to a play or a big sports event online. You’re almost done when that annoying Captcha screen comes up and makes you type some blurry letters and numbers into a box. This step, as most people know, is to ensure that you’re just a person buying tickets and not a computer program deployed to illicitly to grab up a bunch of seats.

But why can’t a computer that can perform calculations astronomically faster than humans identify the letter B just because it’s in a fancy font with a strikethrough, or the number 5 in a fuzzy photo of a front door? Why is it so easily baffled by something the average second-grader can handle?

The answer lies in understanding the current state of artificial intelligence (AI) — what it’s capable of, what is still beyond its grasp, and how we may be rocketing toward an increasingly intelligent technology without enough thought about the implications for ourselves and our planet. This is according to Tim Urban, author of the quirky, stick-figure-illustrated, popular blog Wait But Why, which counts Tesla CEO Elon Musk and Facebook CEO Mark Zuckerberg among its fans. He spoke recently at the McNulty Leadership Program’s Authors@Wharton speaker series.  .... "

Google Rethinks the Captcha

From CWorld: In attempt to get away from the often annoying Captcha interaction to prove you are not a robot.  With a far simpler subtle approach. Will this be enough for the more clever robots of the future?   I assure you this blog is written by a human.

Spam is Improving AI

In the sense that research underway to solve CAPTCHAs is leading to improvement in some kinds of pattern analysis. Ironic AI challenge between those that create the Captcha letter puzzles and those that solve them. Is this an argument for gaming in business? Maybe it is not about competing to find a better solution, but competing at contrary aspects of the same problem? Can a business problem be positioned that way?

An Intelligence Economy

Telling article about solving Captcha puzzles in India.

Games With a Purpose

I have been looking at a number of online 'games with a purpose' aka 'serious games' , especially how they can be related to more elaborate simulations ... Here is something new and related:

" ... Luis von Ahn and his team at Carnegie Mellon University have launched GWAP, a new web site for 'Games With A Purpose.' By playing these online games, humans help provide data for problems that are hard for computers to solve, such as computer vision and sound classification. Slashdot has previously covered other human computation projects by Dr. von Ahn, including the ESP Game and reCAPTCHA. The new web site contains a re-vamping of the ESP Game as well as four completely new games." ...

Mentioned in the BBC. Note this is about getting data from people, somewhat different from playing complex games to achieve process solutions, but there is a useful connection.

Captchas and Ann Hathaway in Encyclopedias

I mentioned the term Captcha in a recent post. The term refers to the distorted letters you are asked to interpret when you sign up for a new web service. The object of this is to detect that there is a human asking for the service, as opposed to a spam-spewing robot. The term is defined in a detailed article in the Wikipedia. It is not to be found at all in the premium version of the Britannica. Yes, I know this comparison may be unfair, the term is a somewhat obscure, fairly new (coined in 2000 the article says) technical term, and the WP excels at this sort of thing.

Then there followed another, non technology, example. Britannica just sent out a set of topic articles, constructed from BT entries. One was about Ann Hathaway, Shakespeare's wife. Exactly 303 words long. I am a Shakespeare buff, and know there is relatively little known about her except her family ties. No technology here. Then I looked Hathaway up in the WP. Much longer article there, lots of links to recent related books and articles, including Germaine Greer's 2007 book: Shakespeare's Wife. By my non-historian analysis, a similar level of credibility about the few known and and some inferred facts. Nothing controversial I could see, writing clarity good. So there too, the WP won in overall value to someone that wants a credible but detailed overview.

Once more, though, I don't want a winner, I want competing resources.

Captchas and ALIPR

I wrote recently about the hacker cracking of Captchas, those distorted letter problems you see when you sign up for new online services.

This is very serious since it has allowed the release of millions of spam messages from bots posing as people. I have also been looking at methods for identifying things in arbitrary images and using them to tag pictures with identifying descriptors.

One of the most interesting approaches to this latter problem was developed at Penn State and is called ALIPR (Automatic Linguistic Indexing of Pictures) It's a fascinating mathematical programming method that can assign sets of tags to images. It turns out that this problem is one of the most difficult immediately practical AI problems known. It is also the case that the ALIPR approach to tagging photos is not perfect, where any particular tagging of a photo includes wrong tags. A recent Cyber Cynic blog post describes how the ALIPR method is being reapplied to the now broken Captcha method, with links to demonstrations of the method. For those with the math background, this paper gives a perceptive view of the ALIPR method.

Gmail Busted, Humanness Suspected

I noticed a few weeks ago that some of my Gmail messages were being lost. A few were kicked back to me labeled Spam. And a few correspondents reported that my messages were being placed in their spam folder. Of course some people, like me, rarely look at their Spam folder, so they never saw my messages. Unexpected mail being lost. Disconcerting to say the least. Snail mail was becoming more reliable. Should I abandon Gmail? I searched around to figure out why. It turns out the the Captcha which Google uses to test to see if you are human when you apply for a Gmail account had been broken. So Spammers were getting Gmail accounts, and creating a new deluge of Spam. Spam filters were reacting by blocking some kinds of Gmail. For example a test showed that a Gmail message containing any kind of URL was more likely to be blocked than one without. Guess I have to wait for a solution. Better tests for human-ness are now needed.

Be Identified

Now that I am responsible for my own infrastructure I am taking more notice of the risk of using specific methods. I was quite surprised, after changing my hardware, to start being challenged by Google's blogger. Many of you have probably seen the challenge, where you have to read a distorted set of letters, to prove you are not a sim, This is called a captcha. Last night this happened several times in a row. The letters I had to read were so distorted I had to try several times. I felt somewhat put off by this. Am I a machine or human? Am I depending too much on Google's usually excellent services?