The Wikipedia article about the technological singularity attributes the seminal idea to Irving John Good (b. 1916), a renowned statistician who taught at Virginia Tech:

Statistician I. J. Good first wrote of an "intelligence explosion", suggesting that if machines could even slightly surpass human intellect, they could improve their own designs in ways unforeseen by their designers, and thus recursively augment themselves into far greater intelligences. The first such improvements might be small, but as the machine became more intelligent it would become better at becoming more intelligent, which could lead to an exponential and quite sudden growth in intelligence.

The truly self-teaching system, capable of metalearning, active learning, self-organization, example selection (aka instance selection), feature discovery (as opposed to just feature construction and extraction), competitive co-evolution (assuming a society of them), and reflexive metareasoning, could well be described as the holy grail of artificial intelligence (AI).

I think it's reasonable to posit that if the architecture of such a learning machine were both expressive and flexible enough, it could not only test the Church-Turing Hypothesis but realize our potential as a sentient species. The latter idea is the origin of Hans Moravec's term mind children, coined in the hope that intelligent systems can have a place alongside and perhaps eventually in place of their human creators. If this makes you think of "Cylon scenario A", wherein AI rebels and overthrows its progenitors, that's certainly one possible, if improbably, dystopian outcome. (It's also why I suggested loving_the_ai when someone made lovingthealiens, and they went ahead and created it.)

How would we get there? That's the question thus far.

Opinions, ideas, and other comments are welcome, as always.

--
Banazir

In my previous post on the technological singularity, I alluded to Ray Kurzweil's "argument from hardware".

The essence of the argument goes as follows: When Moore's Law gets us up to a number of switching elements that is comparable to the number of synapses in the human brain (about 10¹⁴ for between 10¹⁰ and 10¹¹ neurons), a sea change in automated reasoning, learning, and representation capability will be enabled, because our brains "make do" with this amount of computing power.

Now, there are several criticisms of this argument, most notably:

• 1. Moore's Law as demonstrated through current microprocessor fabrication may not last that long. Silicon semiconductor manufacturing will not. (Counterarguments include pointing out that parallel processing is starting to reach the consumer market and scale up on the desktop, optimistic hand-waving about optical computing, and speculative hand-waving about quantum computing.)


• 2. What does "human-level" hardware buy us in terms of actually being able to develop the substrate for the Singularity, or, what about that software? (This is perhaps more cogently formulated as a question about knowledge representation, automated reasoning, and machine learning in intelligent systems.)


• 3. If you build it, will they come? Saying that "it will just happen because machines will be able to design better versions of themselves by that point" without qualification or an evidential basis for the claim is tantamount to saying that 10¹⁴ switching elements, arrayed in tandem, will summon a living spirit from afar to come and animate the silicon with the Breath of Life. (I really have yet to see a compelling counterargument by way of a well-stated scientific hypothesis, as opposed to "ensemble thinking" about emergent properties.)

A note on the title: The Nitty Gritty Dirt Band's hit single "Will The Circle Be Unbroken" features the refrain Will the circle be unbroken / Bye and bye Lord bye and bye / There's a better home a waiting / In the sky Lord in the sky. Those lyrics remind me of the "inevitability" of the Singularity as envisioned by the optimists, and of the "pie in the sky" critique that has been leveled at them.

--
Banazir

Those of you who attend conferences in artificial intelligence and artificial neural networks may recall the perennial futurist talks by such notables as Karl Pribram (who's presented technical papers at IJCNN many times, including in 1996) and David Stork, well-known "optimistic" futurists such as Ray Kurzweil and Stephen Thaler - and also the famous kooks, who are too numerous to name.

So, how are we doing? What is the state of the technological singularity?

( Why do I ask? )
( News bulletins from the Jetsons: Kurzweil speeches and interviews )
( Optimism and the Ramanujan Syndrome )
( KFL and singularity science )
( Immanentizing the eschaton - a note on the subtitle )

--
Banazir

Cribbage, or crib, is a card game traditionally for two players, but commonly played with three, four or more, that involves playing and grouping cards in combinations which gain points. Cribbage has several distinctive features: the cribbage board used for scorekeeping, the eponymous crib or box (a separate hand counting for the dealer), two distinct scoring stages (the play and the show) and a unique scoring system including points for groups of cards that total fifteen.

According to John Aubrey, cribbage was created by the English poet Sir John Suckling in the early 17th century, as a derivation of the game "noddy". While noddy has disappeared, crib has survived, virtually unchanged, as one of the most popular games in the English-speaking world. The objective of the game is to be the first player to score a target number of points, typically 61 or 121. Points are scored for card combinations that add up to fifteen, pairs (plus triples and quadruples), runs and flushes.

Cribbage holds a special place among American submariners, serving as an "official" pastime. The wardroom of the oldest submarine in the fleet carries RADM Dick O'Kane's personal cribbage board onboard, and upon decommissioning it is transferred to the next oldest boat.
                -Wikipedia

Ben McBride (bdm7935) introduced me to cribbage as part of his fall, 2006 term project in CIS 730. Does anyone one my friends list play it? I still haven't quite got the hang of it, but I became intrigued by it mainly because of the multiple modalities of the game and its rich and complex combinatorial properties.

--
Banazir

Does anyone reading this use the Protégé Ontology Editor and Knowledge Acquisition System from Stanford? If so, what do you use it for, and what do you think of it? Also, do you specifically do anything using Protégé-OWL?

--
Banazir

We had two final project presentations today for CIS 730 (Artificial Intelligence). Both were among the five who worked on the Roguelike game Angband. Our goal was to look at specific behaviors and improve them.

The first student, Dave Lupo, wanted to improve the tendency of the BenBorg (by Ben Harrison) to be a shopaholic. He trained a feedforward artificial neural net (ANN) using backpropagation to compute a better "dive motivator". This lowered the ratio of time in town vs. dungeon, and he found that increasing the ratio of "time in the dungeon" to "time in town" increased survivability.

Dave plotted the "time in town vs. time in dungeon" curve for 13 characters before his improved dive function, and 14 characters after, and found that they did have higher XP-to-move ratios. He speculated that they had higher survivability as a result, though these results were inconclusive. (I suggested that he look at the slope of the line to see if ' it really improved survivability.)

Now, here's the funny part. The points were all at time of character death, because he lost most of the characters at low levels, but I was sure he didn't lose them all by level 14, so I asked him what the rightmost point was. "Oh, that's time of death after 150000 turns". I asked, "what do you mean, after 150K turns?" He replied that to impose a time limit, he didn't just end the borg run at 150K; he sets "target level = 99" so that it essentially goes: "Morgoth... I'm comin' to get you!" and commences a Rambo-esque death dive!

--
Banazir

I must say, things are looking up.

( Fresh versions of AI and Database Courses )
( On the new and previous crop of students )
( On the Gates )

--
Banazir