Month: September 2009

“My laptop’s so old…”
“How old is it..?”
“My laptop’s so old, three undergrads have come up to me and said: wow, cool laptop!”

Today is our department’s annual retreat in Leavenworth. Three weeks ago, knowing
my car’s failing radiator would not sustain the 120 mile journey, mostly on mountainous roads, I asked for a seat in one of the department vanpools. This preemptive measure set my mind at ease.

Last night my buddy said that he would appreciate my company on his drive up.  “Win for me” I thought, rather than spending three hours in cramped confines having to behave myself in front of respectable professorial types, I’d get to enjoy good music, raunchy jokes and juicy gossip in my buddy’s car.

This morning, I woke up on time, packed, grabbed some breakfast and drove to his place.  The plan, as I understood it, was to depart at 9 am and then pick up another friend of ours on the way out.  On arrival I discovered his car mysteriously absent and no answer at his door.  I waited about ten minutes then called his cell phone and left a message. Panic struck when my phone flashed 9:15.  Having no confidence in my memory, I figured I mixed up our agreed upon meeting time, my buddy got fed up waiting and headed out without me. 

The vans had already left, and being on the hook for several duties for the event, driving my jalopy seemed the only course of action left available.  With a tank full of gas and a trunk full of anti-freeze, I said prayer to Bitchasschles, the god of perilous- (formerly snow-) driving, and set off east.

Having just rounded Lake Washington my car’s idiot gauge suddenly climbed to “H”.  I stopped in a turnout, grabbed a gallon of coolant (one of 5) from my trunk and dumped it down my parched radiator—I would have suffered steam burns had it actually contained fluid.

Back on the road, I get a call from my buddy:

“Where the fuck are you!?!” 

“Uh, I was by your place earlier and you weren’t there…” I
replied meekly.

“When?”

“9:05, and I waited ten minutes.”

“I was getting gas.
We were supposed to head out at 9:15.”

“I left a message on your cell phone.”

“I never check my cell phone; you know that.”

“I thought the plan was to pick up Anne at 9:15?”

“No.  We were to
meet her at Ly’s doughnuts at 9:30.”

“Well, shit… sorry, I guess.”

“You’re hopeless.” [click]

I continued the drive on highway 2, my heater on full blast, studying the oscillating needle of my temperature gauge.  I checked the road occasionally so I wouldn’t slam into the occasional RV.

After fifth time I stopped to replenish my radiator, I turned on the engine and studied how, what had been a drip of fluorescent green 2 hours earlier, had become a steady stream–my car relieving itself alongside of the road.

I coasted close to twenty miles down Stephen’s pass with my engine off, which bought me some miles before the my next stop. 

Bitchasschles came through for me, once again.  I arrived in Leavenworth, on time, and my car still running though my engine likely sustained some overheating damage.  My buddy and I made up.  He acknowledged he was a bit cranky since he spent all but 3 hours last night working on his poster.  With the exception of twisting my knee playing volleyball, I’ve been enjoying the retreat so far.

Artificial neural networks are computer programs that simulate how a collection of representative neurons (or more accurately: groups of neurons) can interact to perform a computational task.  Neural networks have been trained to recognize faces, read handwriting, complete partially-presented patterns and continue temporal sequences–in short, they perform many of the tasks that our brains can perform with ease, that computers, programmed with conventional rules-based routines, cannot.

In the 80’s and 90’s, the Department of Defense funded most academic research on artificial neural networks.  I speculate the military hoped for a means to imbue machines with cognition.  Perhaps so they could make themselves some Terminators or whatever.  These days, interest in artificial neural networks has waned as many cognitive scientists contend artificial neural network models do not adequately capture the more intricate physiological mechanisms underlying cognition.  Consequently, many have deemed artificial neural network models unworthy of study. 

Few would dispute that neural networks are very crude models of how the brain could perform a cognitive task.  It is the physicist in me that finds them fascinating and worthy of study.  Like sub-atomic particles that comprise matter, neural networks are more than just the sum of their parts.  Their smallest elements, units, are very simple–each takes input from, and sends output to, the other units in the network.  The connections between units vary in strength, but each unit simply adds together all of the input it receives (from the other units in the network) and conveys this sum with its output.  That’s all.  To first order, this is what physiological neurons do as well.

The magic of artificial neural networks lies in the connections between units.  A connection dictates the extent to which one unit’s output activity affects another unit.  It is how these connection strengths are adjusted that differentiates a network that, say, recognizes faces from one that generates a sine wave when cued.  Though these examples perform very different tasks, their underlying functional element, the unit, is identical.  Given enough units and the processing power to appropriately adjust the connections between them, artificial neural networks can be trained on a great variety of complex cognitive tasks.
 
Could neural networks ever embody an artificial intelligence?  Probably not.  But my materialist philosophy leads me to believe it is the right track.