AI Panic!

I found a wonderful comic while procrastinating socializing on Google Plus. Please click through to the image to see the full comic, made by Doghouse Diaries.

Having an Android-phone myself, I haven’t really used Siri much, but the speech and context recognition capabilities are quite impressive. Don’t make a mistake - there is no “sentient” being in your phone. While the algorithms used can clearly be called AI algorithms, but mostly it’s relatively straight-forward search. Siri (and related chat bots) use a large database to look for context clues (like “date” “meeting” “remind” “reminder” and so on for a calendar entry) and then use grammar rules to deduct the corresponding data (date and message for the calendar example).

There is a long debate about how machines can really understand human language (and what understand actually means in this context). Direct commands (Siri) and questions (Wolfram Alpha and Google) are a bit easier to decipher - however they still have their issues.

Sadly, John McCarthy has passed away yesterday, at the age of 84.

McCarthy is one of the founding fathers of AI, he organized the important Dartmouth Conference, which is now considered as the defining event that pretty much created AI as a field (John McCarthy himself coined the term “Artificial Intelligence”).

As a mathematician, he championed the use of logic in AI, and invented the programming language Lisp in 1958, a multi-paradigm high-level language that quickly became the language of choice for artificial intelligence research. I found this nice and sad programming analogy earlier today on reddit:

)))))))))))))))))))))))))))))))))))))))))))))))))))))))))))

Anyone who codes in Lisp knows what it means to type a long series of close parentheses — the more parentheses, the bigger the work being brought to a close.

Unfortunately, in real life, we seldom get the opportunity to close all of those parens. Life just ends with a whole bunch of unmatched open parentheses.

In his later years, McCarthy became very interested in the sustainability of progress and looked into the question of simulated emotions (”The Robot and the Baby” is a lighthearted short story he wrote on the topic).

It is a sad day for us, having lost our proverbial father. In fact, it has been a sad month, with Dennis Ritchie (the inventor of C and co-creator of UNIX) also passing away.

Actually I didn’t know that this record wasn’t already in the hands of a machine, but now it surely is: solving a Rubik’s cube as fast as possible. Mike Dobson and David Gilday built a robot called CubeStormer II, using sets of LEGO MINDSTORMS NXT and a Samsung GALAXY S II smartphone. Have a look at the video:

While definitely easier than understanding Go, the challenges here are different, and the solution is remarkable for a couple of reasons:

• Image Recognition: The state of the cube has to be recorded and recognized from an image. Everyone familiar with image recognition knows that this is a very error-prone task and there are hoards of PhD students and researchers working on recognizing faces and objects in pictures. That said, recognizing the colors of squares on a regular cube is not too bad, especially with good and even lighting.
• Finding an optimized solution, quickly: This is one of the easier challenges, as the state-space of a Rubik’s cube is not too big, and a path to a solution can be found reasonably quickly. While it might require something a bit faster than brute-force, there are plenty of fast algorithms out there. Still, it needs to run in milliseconds and convert the solution into a series of motor commands for the machine.
• Fast and accurate motor control: Solving a Rubik’s cube very quickly requires (apart from a well-lubricated cube) very fast and accurate control of the layers. Once the robotic grip of them slips or fails to align perfectly, it’s basically game over, and would probably also lead to destruction of the mechanics.
• Availability of parts: The robot was built using a mobile phone and Lego, both of which are readily available. While it might use some custom parts to optimise for speed, nothing would stop you to re-build a similar robot using your smartphone and your Lego-set. I’ve built a small, pointless robot using an Arduino and Lego myself, and it’s easy and good fun for a lazy weekend!

The computing power in modern smart-phones is incredible, and combined with cheap mechanical sets such as the Lego Mindstorm NXT, which can be found in many kids rooms these days, robots can be built that are better at their tasks than any human. This trend will continue and accelerate, and will be used for good and evil, I’m sure (Panic Level: +1%). While this might give reason to worry about the safety of kids playing with extremely computationally powerful toys, I believe it is crucial to grow up playing and experimenting with them, as an increase in automatisation and “robotification” seems inevitable at this point.

There are plenty of websites that keep on top of the hobbyist efforts, such as Hackaday.com.

(via SingularityHub)