Channels ▼

Cameron and Tracey Hughes

Dr. Dobb's Bloggers

One Little Question, That's All...

March 16, 2009

After all, what does Project Purity have to do with Megaton and Vault 101? Now that you mention it, who watches the watchmen? How will we know when the Singularity is near? All important questions in some domain. Each question hiding an even more fundamental problem.Make no mistake about it! Depending on how the question is formulated, the problem and any potential solution might be forever hidden from sight. On the other hand, if the problem is in plain view then, with a little luck, we might ask the right question. And if we ask the right question then we at least have a chance at finding an acceptable answer.

It's no secret that Tracey and I use the computer as primarily a problem solving device. Again, not as some integral part of a solution but as the principal problem solver. We end up with all of these hard problems to solve and we then set out to engage the computer's problem solving capabilities. Almost invariably the difficulty of the computer's task is tied directly to how we phrased the question and indirectly to what we really expect the answer to look like. See, it's deceptively simple. We now have at our disposal incredible computing power. We have on our heels some pretty tough problems. If we could just bring to bear our incredible computing power on our tough problems, the solutions should just present themselves right? Our computers now have the capacity to hold trillions, quadrillions, even quintillions of bits of information (duodecillion capacity on the way). No longer are we restricted to single CPU work horses. We now have multicore processors that can execute billions of instructions or trillions of floating point operations in an instant. We can easily tie multicore computers into clusters. We can then combine clusters into supercomputers. What problems can withstand such computing power? The silver lining behind our biggest clouds is nothing short of massive parallelism. So with all of that star power, Tracey and I pose the simplest questions to the agents that inhabit our multicore magical lamps, and we find out first hand what relation needle-in-haystack has to AI-complete. The heat emanating from our supercooled multicore processors makes it very clear just how serious the search for answers can be.

Its widely known that the Internet contains the sum total of human knowledge, right? Certainly if there is anything that is known it can be found somewhere on the Internet. Using the Internet as one massive knowledgebase, that's what the Semantic Web is all about, right? Sure the Internet is big. But with a powerful enough computer configuration I should be able to blitz through it and find my answers, right? To get some sense of the deceptively simple problems Tracey and I now give our parallel processing partners, here is a typical question that we might present our agents with:

Which tastes better, coke or pepsi?

How large is the search space for any agent(s) that attempts to answer this question? How long should it take for our super charged multicore machines to find the answer? What does this trivial question have to do with AI-complete problems? Perhaps even trickier than finding the answer to this, is determining how long should it take our super charged computers to understand the question. No keyword matching here!, No probability indices! What we need is good ole fashion understanding of the question that is posed. What should the answer look like? Should it contain any qualifiers, quantifiers, caveats? If the agent has access to the entire Internet as a knowledge base, how large is the search space for understanding the question and finding the answer? We've been playing with this for a couple of months now. No matter how we look at it, we're coming up with astronomical numbers. In the words of Rachel Maddow, can somebody please talk us down? Is there some optimal number of parallel search processes that would make the agent's task any easier? By the way, did we fail to mention that the coke and pepsi we are talking about were two chickens that we raised when we were young?

Related Reading


More Insights






Currently we allow the following HTML tags in comments:

Single tags

These tags can be used alone and don't need an ending tag.

<br> Defines a single line break

<hr> Defines a horizontal line

Matching tags

These require an ending tag - e.g. <i>italic text</i>

<a> Defines an anchor

<b> Defines bold text

<big> Defines big text

<blockquote> Defines a long quotation

<caption> Defines a table caption

<cite> Defines a citation

<code> Defines computer code text

<em> Defines emphasized text

<fieldset> Defines a border around elements in a form

<h1> This is heading 1

<h2> This is heading 2

<h3> This is heading 3

<h4> This is heading 4

<h5> This is heading 5

<h6> This is heading 6

<i> Defines italic text

<p> Defines a paragraph

<pre> Defines preformatted text

<q> Defines a short quotation

<samp> Defines sample computer code text

<small> Defines small text

<span> Defines a section in a document

<s> Defines strikethrough text

<strike> Defines strikethrough text

<strong> Defines strong text

<sub> Defines subscripted text

<sup> Defines superscripted text

<u> Defines underlined text

Dr. Dobb's encourages readers to engage in spirited, healthy debate, including taking us to task. However, Dr. Dobb's moderates all comments posted to our site, and reserves the right to modify or remove any content that it determines to be derogatory, offensive, inflammatory, vulgar, irrelevant/off-topic, racist or obvious marketing or spam. Dr. Dobb's further reserves the right to disable the profile of any commenter participating in said activities.

 
Disqus Tips To upload an avatar photo, first complete your Disqus profile. | View the list of supported HTML tags you can use to style comments. | Please read our commenting policy.
 


Video