Sunday, December 08, 2013

Lessons from Android Cephalopods


I’ve been thinking a lot about robots recently.  
It started back in September, when I read an article in Scientific American where the author posited a time in the near future where our relationship with robots would shift.  Currently robots are employed as servants, given tasks too dangers or too boring, or both, for humans to do.  Military drones and automobile assembly lines are good examples.  Others are coming to do things like explore the rubble of collapsed buildings after disasters and to locate and dispose of hazardous waste.  
But the article I read talked about how, in the very near future, that relationship would change to where human employees would take instructions from robots and AI expert systems.  These constructs would direct their human counterparts in assisting them to get work done.  An example was cited of a robot welder that had been created.  Working with a human assistant, to move and set-up the parts to be welded based on the robot’s commands, it was able to weld a frame for an armored vehicle much faster and less expensively than a team of four expert human welders.  
The article struck a cord in me.  As a manager of a production department, with about fifteen reporting to me, I had sometimes joked (during some days of extreme frustration) that I would replace my entire staff with robots if I could.  I would use a small force of robots with the capacity for telepresence.  They would work the copiers, move the jobs from one station to another, while an expert system would review the documents for legibility and accuracy.  When a problem came up that the AI system or the robots couldn’t handle I would get a message, remotely access the unit with the problem and fix it, using my solution to train the robot or AI in how to solve such a problem in the future.  I even had the model that I would want to buy, Right Here.  I figured about five of them, working twenty-four hours a day, would be able to replace my entire staff.  
What this article said to me, though, was that if anyone was likely to be replaced in such a work situation within the next twenty to thirty years, it would be me.  Having an AI, programed with an understanding of legal requirements for legal document production would be relatively simple, in time.  And the facets of the work that I saw myself teaching a robot staff, such as character recognition of hand-written doctor’s notes, could be handled by a human assistant under the instruction of a robot production manager.  
Damn.  Guess you should be careful what you wish for.  
It did inspire me to write a story about a Robot Boss.  It was the same story that nearly made me decide to give up writing all together.  I finished the rough draft for the story this week.  The key to finishing it was the idea of what the human assistant teaches his robot boss, and how human perception differs from that of robots (at least now and in the near future).  
Learning was also a part of another article about robots recently.  This one was about the efforts to create a robot octopus.  
Octopuses (octopi?) are the smartest invertebrate on the planet.  Their bodies have capacities that are envied by people in the field of robotics.  The soft body of a full grown octopus can squeeze itself through a hole the size of a quarter.  It can camouflage itself, changing colors to match its environment almost instantly, as well as shoot out a cloud of ink to disguise its escape as it jets away.  Robots designed by humans to work underwater have been rigid affairs, with limited motion and manipulation capacity when compared to these creatures.  
This article, which also appeared in a recent issue of Scientific American, underscored one of the reasons octopi (octopuses?) are so smart despite having limited brain capacity.  I remember reading years ago that octopi or ~puses have been rated as being about as smart as a dog, though their brains are only about the size of a walnut.  The reason for this apparent discrepancy is because the animal’s tentacles essentially control themselves.  The scientists referred to this as “embodied intelligence.”  The tiny octopus brain gives direction to the tentacles (“Grab that lobster so I can eat it!”) and the tentacles take care of the gross movement necessary to do so.  The neuroscientists associated with the robot octopus project didn’t know how the neurons within the tentacles did this.  The people working on designing the robot octopus, though, thought they could use the same concept to get their robotic tentacles to learn how to move and manipulate the objects around it.  It would be simpler than normal robotic programing, in fact.  Instead of programing the tentacles with a catalogue of specific motions to use at different times, they would only have to give them a small handful (or tentacle full) of motions, which would then be adapted as needed to overcome any problem the artificial cephalopod might encounter.  
I’ve started wondering if there’s a story about working for a Robot Octopus Boss.  Nothing has come to mind yet.  But the idea is intriguing.  In a weird sort of way.
Shifting gears a bit...  In recent weeks, when it comes to my writing, I feel like I’m in rehab.  Last month, as indicated by my last two blog entries, I was in a pretty extreme state.  I didn’t want to give up writing per se, but I wanted the frustration and the sense of wasting my time to stop.  And giving up writing seemed to be the way to do it.  
I don’t feel that way now, but I do notice that things are different.  My attitude and my approach to what I’m writing is different.  At times, it almost feels like something is missing.  I still get up at the same time, for the most part.  And I’m close to putting in about the same word count and hours.  But the...  Drive, maybe?  Or the insistence that I get “something done”...?  I haven’t identified what it is that’s different, just that something is.  
Maybe it’s not important.  It seems to be something about my writing and not within my writing itself. Maybe my writing really has a life of its own.  A part of me that is in control of itself the way the octopus’s...  Octopi’s...  That aforementioned cephalopod’s tentacle’s embodied intelligence directs its own motions.  My job, like the tiny brain inside that head designed for an alien mastermind, is to simply learn how to tell it what I want it to do.  
Maybe, for robots and people, it all comes down to learning.    

0 Comments:

Post a Comment

<< Home