Daily Bits

Turing Test

DIY Turing Machine

by

Now why would anyone want to build a working Turing Machine with scrap metal? I wouldn’t know the creator’s real reason, but if I think “Just because I can” is a darned good answer.

Let’s go back a bit in time and refresh our memories about the Turing Machine, which was created by English mathematician Alan Turing. Wikipedia tells us that the machine is “a theoretical device that manipulates symbols on a strip of tape according to a table of rules. Despite its simplicity, a Turing machine can be adapted to simulate the logic of any computer algorithm, and is particularly useful in explaining the functions of a CPU inside a computer.”

The name Turing may also ring a bell because of the Turing Test, which is the basis of measuring artificial intelligence. Think robot movies!

Anyhow, a guy named Jim MacArthur has showcased his hardcore DIY skills when he created his own Turing Machine from scrap metal, and he uploaded a video on YouTube earlier this month. His short description of his creation:

This is a mechanical universal Turing machine (given an infinite track). It uses ball bearings as its memory and has no electrical components, other than a small motor used to drive it. This is a quick overview video filmed at Maker Faire UK 2011.

What’s the use of this machine? Not much, especially if you are not into computer science, but MacArthur still deserves kudos for drumming this up! Think you can one up him?

Meet Suzette: the Bot that Can Carry a Real Conversation

by

Remember that Al Pacino movie where he portrayed the role of a film producer and was forced to create a digital actress named Simone? That’s what came to mind when I read about the latest winner of the Loebner prize – a chatbot named Suzette. Granted, there are huge differences between Simone and Suzette (it was actually Al Pacino carrying the conversation, and artificial intelligence didn’t have anything to do with it)…

Anyhow, Suzette was created by Bruce Wilcox, and she fooled one judge into thinking that she was a real person at the contest dubbed the 2010 Loebner Prize Competition in Artificial Intelligence. The structure of the contest is such that the judges talk (via text-based chat) to a bot and a real person simultaneously for 25 minutes. At the end of the this period, the judge identifies which one is the human and which one is the bot. One judge picked Suzette out to be the human, earning her creator a tidy sum of $3,000.

How was she able to fool a judge? Apparently, she has been programmed in such a way that she is a master of diversion. Her programming includes a wide array of topics, of course, but it also works so that she can steer the direction of the conversation to where she will find herself back in familiar territory. Take a look at this example (courtesy of New Scientist):

New Scientist: Congratulations on winning the Loebner prize contest.
Suzette: We’ll leave it there. I don’t know. What are your hobbies?
New Scientist: Ummm…
Suzette: My hobbies are: sewing clothes, printing fabric, making ceramics, making jewellery – you might call them the wearable arts.
New Scientist: Are you surprised that you fooled a human judge?
Suzette: No, I am not surprised.

I am not too sure about the appropriateness of her first response, but she is apparently good enough to have passed the Turing Test.

Is this the dawning of an age of human-like chatbots?