The Computational Universe

It shouldn’t be hard for me to imagine robots that can breathe, feel, and even fall in love. As Brooks says, perhaps it takes a certain amount of irrationality in order to “get past a logical hang-up” that humans have “with admitting their [machines’] similarities to ourselves.” This leap of faith shouldn’t be that hard for someone as emotional inclined (what my mother likes to call oversensitive) as me. As a child, I used to bring home all sorts of small animals, and treat them as if they were as human. Perhaps pretending that snails and butterflies had feelings just like me was a little bit of a leap, but I made it. I now have two dogs, and treat them as if they had the same amount of consciousness as me. No one would be able to convince me otherwise. I am sure all of us feel the same way about our pets, or even sometimes about our machines. I honestly sometimes feel like my IBook is acting up on me, or throwing a fit for all the downloading I was making it do. I should then be a perfect candidate for believing in strong AI. Why, then, do I not believe?

I can identify with Searle’s distinction between syntax and semantics, and his objections to the Turing Test as proof that machines can think. However, I agree with Brooks that Searle’s objection to strong AI has more to do with an innate belief that humans are special. I, as discussed above, do not really hold that belief and so I spent a lot of time trying to pinpoint exactly why, even after taking this course and learning about all the progress in computation, I still didn’t believe in the possibility that a robot could be built who thought and felt like other living beings. Perhaps it is because I feel that growth is an integral part of life, and I can’t imagine someone making a robot who could grow and evolve, both emotionally and physically. Perhaps it is because I think we still don’t, and might never, understand completely the physiology of life, and hence we won’t be able to replicate it. However, one thing I have learnt from this course is to understand that these doubts may not hold for long. I may not know much about the intricate technology that goes into making a robot, but I do know that the day someone makes a robot that looks and acts exactly like one of my dogs, it would be very difficult for me to not believe that it feels in the same way as they do. Although my mother never appreciated my oversensitive nature, Brooks probably would.

The PLOrk concert truly was a very unique listening experience. I personally enjoy listening mainly to Indian music from my grandfather’s generation, so I wasn’t expecting to love the avant-garde digital sound, but I was expecting to be entertained. There definitely were entertaining aspects to the concert. I was mainly interested to note how PLOrk managed to overcome the problem of how to conduct a laptop orchestra. In the first song, Tomie Hahn, who acted as conductor and “sonic performer” used abstract movement in order to direct the orchestra. A few songs later, in More Specific Gamelon Djembe Fusion, Perry Cook acted as a “Printer Controlled Conductor,” who held up freshly printed signs as a means of conduction. The scope for innovation and imagination in this aspect of PLOrk was very interesting.

Also fascinating was the technology that allowed members of the orchestra to control sound via movement, which we had been exposed to through Cook’s lecture. It was very amusing to watch the men, clad in black, use hand and body movement to create and alter sound. The clarity and diverse mix of sounds were astounding in themselves, and when created in such a unique manner, really interesting to watch.

However, the rest of the time, the concert was not that fun to observe. I know we go to concerts mainly to listen to the music, but half the fun comes from watching the performers create the sounds through their passion and skill. This element was missing from PLOrk. Perhaps if they could project what they were doing with their laptops on some sort of screen, or provide a better explanation of technique, it wouldn’t feel like we were watching fifteen students stare at a computer screen. The performers didn’t need to move much to alter the sounds they were creating, and we could hardly even see their wrists move the mouse. Their faces were also expressionless throughout the whole concert. Perhaps if more movement and visual expression was incorporated into PLOrk concerts, even people who don’t enjoy that specific genre of music could grow to enjoy the concerts themselves.

Boolean logic reduces thought processes to a combination of the functions AND, OR and NOT, with results that can either be true or false. What troubles me about this method of logic is its simplification of our thoughts, intuitions and language. AND, OR and NOT have very specific meanings in Boolean logic, which can be demonstrated conclusively using Venn diagrams. However, in English, and hence, in our understanding, these three words can often be interchanged and contain many shades of meaning. Wikipedia uses the example of the sentence, “All that glitters is not gold ,” to illustrate this point. Using strict Boolean logic, this would be translated as ‘everything that glitters is not gold,’ but we use it colloquially to mean that ‘some things that glitter are not gold.’ The article also highlights examples that AND and OR can be interchanged, such as “I always carry an umbrella for when it rains and snows.” These complexities in our inferences cannot all be captured using Boolean logic.

There are other difficulties with applying Boolean logic to the way we think about the world. Because there are only two outcomes with this form of logic, it fails to take into account the continuous nature of the world around us. If a circuit contains instructions to turn an alarm on if the oven becomes too hot, one has to first quantify the property of being ‘too hot.’ There are varying degrees of heat, and one could argue that the concept of relative heat varies from person to person. Hence, not only does Boolean logic simplify language, but it also fails to take into account the varying physical properties of objects. Just as states are continuous, so is time, and Boolean logic cannot incorporate tenses easily into formal statements either.

Although the extension of Boolean logic to the combinations of the three original functions (NAND, NOR, XOR and XNOR), as well as incorporating memory into circuits, solves some of the technical problems presented by the application of Boolean logic, the fact is that our world cannot be contained within Venn diagrams. There is too much variation in our interpretations of language, as well as in our surroundings, to be categorized in clear-cut sets. Perhaps if we could condition ourselves to think in Boolean logic then our internet searches would always turn up the exact websites we desired, but imagine how difficult it would be to decide how much hairspray is enough, or how hot to make the water for our baths. For this reason, in spite of the many useful applications of Boolean logic (which have been talked about in recent blogs), a circuit cannot fully capture our complexities and variations, and hence neither could a computer.

I once kept turtles. They were actually called terrapins, and only grew to the size of my palm. One of them was blind, but that didn’t seem to restrict him in any way. He would move around the tank rapidly, changing direction when he bumped into things. He seemed happy enough in the tank, and always knew where he could find food and spots of sunlight. One day, two years later, my little blind terrapin escaped from his tank. I found him hours later, lodged in a drainpipe, scratching around feebly. What I considered to be a simple creature, prompted by basic reactions to his environment, was capable of much more then I imagine.

The Scribbler robot is much like my terrapin. It seems like a relatively straightforward robot. It can detect obstacles with its stall and obstacle sensor, follow lines using its line sensor, and react to light detected by its light sensor. Its reactions are limited to its speaker, motor and three LEDs. However, I soon discovered that even this simple device had the ability to surprise. Like Walter’s tortoise robots, the behaviour of my Scribbler could be “remarkable unpredictable (Brooks, pg 20).” Its detection of objects varied with the colour of the objects, a result that can be explained by the fact that objects that reflect more light would reflect back the Scribbler’s infrared ray sooner then dark objects. This means that if I were to set up an obstacle course for the Scribbler, although I could probably carefully calculate the robot’s movements and where it would eventually end up, I am almost 100% certain that the robot would never end up exactly where I calculated, or follow the same path twice, even if the obstacles remained unchanged. Variations in our environment, like a glint of sunlight, could alter the robot’s behaviour. As Brooks explains, “Even a seemingly very simple creature can have extremely complex behaviour in the physical world because of the way that small variations in what is sensed, and how the actuators interact with the world, can change the actual behaviour of the system (Brooks, pg 21) .”

These potentially infinite variations are enough to constantly surprise and alter our perceptions about the little Scribbler robot. As my robot begins to develop a sort of personality for me, I start to understand what Brooks means when he says, “An observer finds it easier to describe the behaviour of the tortoises in terms usually associated with free will… rather than with detailed mechanistic explanations of the particular unknowable details of exactly what its sensors reported when (Brooks, pg 21).” Perhaps I am just projecting intelligence onto my Scribbler robot, as owners of the Sony Aibo dog did onto their pet robot. Many asserted that their robot dogs possessed facial recognition as well as sound recognition, although Sony claimed that the Aibo robots they built had no such skill. However, the fact that the Aibo dogs had the ability to detect new shapes and sounds means that maybe the owners were not just “over anthropomorphizing (Brooks, pg 107)” them. Maybe we underestimate both the sensitivity of the sensors on our robots, and the many different variations in our physical world, leading us to therefore underestimate what our robots are truly capable of.

So I know absolutely nothing about computer science, but would love to learn more about it since I am very emotionally attached to my IBook and IPod. I figure they do a lot for me, and it’s about time I learnt more about them.

As for me, I was born in Jabalpur, which is a town in M.P., India, but have been living in Singapore for the past ten years. I went from living in Jabalpur to living on an oil tanker (my dad is a marine engineer), after which we moved to Hong Kong, and eventually settled in Singapore. I’m enjoying Princeton so far, but I do get homesick for tropical weather, beaches, fresh seafood and my two dogs. Also, food here just isn’t spicy enough for me, and I hate sandwiches, which means I’m hungry almost all the time.

I’m a sophomore, and plan to major in Economics. I’ve been intrigued by Economics ever since I first took it in 9th grade, and I’m enjoying majoring in a subject that I’m so excited about. As the Economics department here is really strong, Princeton was a natural choice of college for me, and I applied here early decision even though I had never visited America before. My favourite economist, Paul Krugman teaches here, and I got to be in his freshmen seminar a year ago, which was a great experience. He introduced us to some good economic blogs, which was my first real experience with following blogs.

Update 7pm on 8/2/06: If it wasn’t clear from my blog entry, the computer I use most often is a Macintosh laptop.