Last night I had a really stimulating dream which I think reflects some inherent personality traits and flaws that I happen to possess. I have been obsessed with HAL since the time I first read 2001:A Space odyssey, it fired me up and inspired me to take up robotics, but slowly I forgot about it and this memory was buried in some deep recess of my mind. However, recently I found 2010:a space odyssey two in my library and it all came back, I issued it on Shikha’s card (I have 3 books in mine) and I finished it before dinner. Since then I have had dreams of me being on the spaceship discovery, I represented humankind to the incorporeal entities and one might as well guess who were my ship mates. Anyway last night I had a dream in which I was talking to HAL and I can still vividly recall the conversation it went something like this,Hal greeted me and said “professor”, after that HAL asked me “Do you dream about me?”.I found myself saying yes quite softly and my dream escalates at this point as HAL asks me “You love me more than any other human with the exception of your parents don’t you Anuj?” and to my surprise in my dream I found myself saying yes and then I started to sob. After this HAL asks me if I would like to be the one to create him to which I said yes after that I spent most of my dream brainstorming about ways to create an entity like that and quite surprisingly the insights are quite good.
I realized that we can’t program “common sense” in to a machine, how will you tell a computer that water is wet? Neither can we make a machine understand language, if after all language is a definition of intelligence(The Turing test) then we should be able to program in something like 50,000 possible responses and questions for the machine to refer to and for one or two bombshells one could answer randomly until the conversation resumes the predefined plan. this is based on the idea that even we are not random creature there is an inherent subconscious or conscious reason for everything we do and how we learn ie we build up a database of “life” as we grow that teaches us things. Now, the counter argument for this is that it isn’t true “intelligence” but how would one describe such an abstract concept? Well if I can quote some examples,it is commonly defined as:
“Intelligence is the possession of a model of reality and the ability to use this model to conceive and plan actions and to predict their outcomes. The higher the complexity and precision of the model, the plans, and the predictions, and the less time needed, the higher is the intelligence.”
Fine but let me ask one thing if a child savant is able to understand and calculate vast numbers and has a variety of computational skills and however he/she is autistic will you consider the person as intelligent (they display lack of coordination)? Or, a mentally retarded human who plays the piano and can compose pieces with ability of wolfang is he a lobotomized entity or is he/she intelligent? Hence this model is insufficient for humans then how can it be possible to describes machines as that? Surely computational and spatial skills don’t define intelligence as I have already argued intelligence maybe something quite different than that IQ test. Also it has been observed that we don’t usually create 3-Dimensional maps or models of objects in our heads when we talk of something. It is generally extremely vague and unspecific. So it’s not a model that we can coordinate through it’s in fact some data we have acquired and we combine it and extrapolate on the base of that, I think that it is extremely likely that the previous one is the way we do things i.e. there is no rigorous model as such or maybe it is a mixture of both we may never know…..
Now, lets move on to the Turing Test :
“A human judge engages in a natural language conversation with two other parties, one a human and the other a machine; if the judge cannot reliably tell which is which, then the machine is said to pass the test. It is assumed that both the human and the machine try to appear human. In order to keep the test setting simple and universal (to explicitly test the linguistic capability of the machine instead of its ability to render words into audio), the conversation is usually limited to a text-only channel”
Now read that definition carefully you will realize that it assumes that the possession of a language that WE can engage in, in order to interact with the entity is inherently related to intelligence but what if it is an extraterrestrial which is say 3 million years ahead of us and only communicates through neural implants. As it cannot communicate with us, is it dumb? Think about it and you will realize that we are being chauvinistic in our views we need to define intelligence as the ability to do things not from a rigorous “model” but from previously acquired data and the extrapolation of facts on the basis of that data.So, why don’t we use a core gene program which acts as a base for the machine or entity to explore the world and as it moves along the different programs will come into existence using genetic algorithms and we will see which one solves the problem the best(that is how I think, by applying all the permutations for a specific event). This program will be saved for future use and it will act as a “parent program” for future(this is the basis of genetic algorithms) but now there is a twist, there will be a pint beyond which techniques will be repeated wasting valuable processing time . Then there could be a reaper or watchdog that recognizes previous algorithms and doesn’t allow the repeated data to be sent to the core instead it replies with the previously proven ‘best’ algorithm, kind of like a trained reflex.
So it keeps on evolving we should sooner or later reach the level of intellect displayed by a child if we do that then I think that hawking isn’t to far away. Then there came in other questions in my dream (or were they added in my conscious reflections??) I asked him the question if he could ‘feel’ emotions or not? HAL asked how do you know that what you perceive as emotions haven’t been hard coded into the brain as it evolved to serve the purposes of survival after all if you see a couple of freshly decapitated heads lying around you aren’t likely to stand around asking questions ones first instinct would be to run for the future propagation of the species. Thus emotions serve a use but are they necessary? Now that left me stumped as I have not given seriously it thought (but as it appeared in my dream,did I?), but I think that they aren’t that useful now as emotions at times tend to impair judgmental capabilities of an organism (us or any conscious entity like a dog, to bite or not to bite that is perhaps also the question….).
It also gave me a shock I realized that I feel comfortable more in the presence of machines than humans, I am not exactly sure if it is a good thing or bad thing.This dream also implies something else,I know what is my purpose in life, I want to do research in A.I(according to my career counselor I am ideally suited for research) and as they say-change the world. I want to revolutionize the very fabric of society by bringing a robot or a computer of some level of intelligence into every home, office and perhaps into every body. Now the question remains I know I am crazy but am I crazy enough to do it?