Intellectual Adolescence

On my forays into the field of artificial intelligence...
These are uncertain times. They say that adolescence is a time of uncertainty and self doubt and identity crises. And emotional upheaval. How true this is of me now; though i may be past adolescence in age, for i am 19, the adolescent period is a very apt description of my current state of affairs in my work in artificial intelligence. I should be embarrassed to use the words 'my work in artificial intelligence'...i've not even been through university, nor am involved in any research project under some institution. Its just personal, private 'research'. The big terms were just borrowed for sake of compactness of expression.
It is a mystery how exactly my interest in robotics arose. (oh no i sound like i'm shrouding myself in a cloud of mystery) But given the general interest i had in technology and science and making and (more often) breaking stuff, this is probably not so surprising. In the earliest period of this interest, which was during my early primary school years, i came across this book at the library's children's section, which was roughly titled "how to make a robot". It was extremely interesting to me, and spoke of all the aspects of the robot, in a rather funky childlike way. Caricatures of grey boxy robots filled the book pages. After that first encounter, i never saw the book again even as i tried in vain to find it.
I also recall exploring the issue of understanding human speech during those 'early days'. That seemed like a pretty easy thing to tackle then, and i created sentences and broke them up into parts and put them in trees...a strategy that i now know (or am i mistaken) is called 'semantics'. I now know that this was the approach of symbolic AI, and presently feel that breaking a conversation into a semantic tree is too narrow and rigid an approach.
In those days my interests were neither focused nor quite serious, especially in AI. Robotics just felt interesting, so i'd claim my interest in it, and pretty much left it at that. Occasional encounters with pictures or videos of robots on television and sometimes in books would remind me of my interest and fascination with robots. Probably the only serious book i've ever read on robotics until recently, was 'Behaviour-Based Robotics' by Arkin. It explained things like subsumption, and introduced the bottom-up approach to AI, which was to create animal-inspired robots instead of trying to build humanoids straightaway. I dismissed such an approach then as i saw no reason why we had to build robotic versions of cognitively primitive animals in order to move on to building humanoids. Why not do it straightaway?
It was only later (but i don't know when or why) that i somehow realised that intelligence should be viewed as part of a wider context. That context is the dynamic interaction between the individual and everything else (ie. its environment). I felt that much of the research in the field of AI (that was and still is my immediate impression, certainly a biased one as i have never looked at AI research journals or surveyed current AI research seriously) was dedicated to the solution of very specific problems such as visually guided movement, natural language processing, or to play some games or solve some mathematical problems. I felt that produced very specific answers to very specific problems, which failed to generalise to broader AI strategies. In essence, they ended up exercising their own intelligence instead of trying to synthesise it. And the solutions were very mathematical. The whole exercise altogether seems very rigid and too narrow. Whatever the method or philosophy utilised, i grouped these under the umbrella of 'symbolic AI' (i understand this may be a misuse of the term. But for the sake of convenience...) Or perhaps we could use the term GOFAI (Good Old-Fashioned AI) instead. yeah...
So i harboured this belief, together with some unjustified belief that neural networks could be THE medium for creating an intelligent system (well, nature used it successfully, so why not give it a try too? That's my only justification. I think it's a good one.) With these thoughts behind me, i carried on with my life (which was now in secondary school and most of my junior college years). And occasionally i would claim to have a fiery passion for robotics and artificial intelligence, while the truth was i did nothing in that direction except perhaps marvel at robots i came across (the movie ones), or mostly scorning them, for they reflected the pathetic state of AI and robotics in general.
When i enlisted in NS, the mind was dulled by military routine and excessive and often senseless regimentation. But i managed to extricate myself from progressive brain paralysis by focusing on artificial intelligence. This time, i was able to identify some fundamental questions and provide some answers to them. Those questions include: the scope of AI, the meaning of 'intelligence', the methodology of research...
Having covered the past and the present, it is appropriate to address the future. The reason is also due to the 'adolescence' that i am going through right now. I am reaching the junction of crossroads, quite uncertain about my next move. That is why i have slowed down and spent some time writing this reflection, to reduce my uncertainty.
My ponderings over artificial intelligence led me to the conception of robocology (i'm not done conceiving it yet!), then introduced me to Alan Turing as a pioneer of the modern computer architecture, his role in artificial intelligence, and his anticipation of connectionism. A BEAM article on Nv neurons and nets by Wilf Rigter brought up the idea of chaos in electrical circuits and the potential of Mark Tilden's nervous neural networks to exhibit chaotic behaviour, which many natural systems including our brains exhibit too. That brought in the interest in complexity and chaos, and led me to wonder how we could design systems with 'chaotic potential' which we could channel to create intelligent systems. Such a development led me to think about many issues of artificial intelligence, and the thoughts i had were random and not directed towards the resolution of particular issues. They were scribbled on pieces of paper. I treasure them, and eventually typed them in two documents, called Miscellaneous Ramblings on Intelligence (MRI). I took those points and put some order in them and created a set of principles and methodology and put them in MRI 3.
However, i felt that MRI 3 did not adequately address the issues of what 'intelligence' meant, and the role of emotions. With the purchase of the book "Mechanical Bodies, Computational Minds", whose articles i found interesting and could very much agree with, i found myself moving into the realm of the humanities. I knew we had much to learn from biology, nature, ecology and neuroscience. And also psychology. But now, i found myself asking questions of a really broad nature which involved sociology too. And the realisation of robocology needs artists and writers and imaginative people. It just means that things have progressed to such an extent that i find myself unable to tackle many of the questions and issues that have surfaced in my mind, and also many important needs of this project. An attempt to tackle all these alone will at best lead to answers with narrow points of view, likely misinformed and misguided. There are also many answers to the questions i ask, and many sources which can provide insight, but are not known to me.
I need to chart a course for action from here onwards.
I cannot steamroll my way through those questions. I need to look for people whom i could discuss such issues with. People from diverse academic backgrounds who can offer what insights their disciplines have to offer. Then what would be my role in this whole affair?
Discussions on those issues can last forever. Innumerable amounts of words were put together to address each of these things, and also related ones. I could take forever, or go for a fast food answer. But i can accept neither.
Dank...nevermind...here's what i'll do:
I'll look for people who could better tackles those issuessettle what i can to the best of my abilities, put them up in well-written documents for comments, and identify and state my uncertainties. Then i'd seek comments and opinions and help. Comments, opinions, help and discussion. Contructive interaction with diverse people.For a purpose.
All things will not be cast in stone.Intelligence is a dynamic phenomenon. So should its research.Keep it haphazard. Chaotic, and well organised at the same time.Purposeful, but not stubbornly single-minded.Its okay that efforts sometimes seem confused. Just don't become confused.Making good rojak is not an exact science, but we all know when we've tasted good rojak. (bad analogy...i don't eat rojak... =P)Don't think like a boring adult. You WILL tend to as the cumulative time spent on this pursuit increases.
You are attempting to recast AI as a multidisciplinary art form. AI will travel further that way than in its old, traditional rigid, narrow and mechanical form.
When i get my rough conceptual framework done, i'll put it out for all to poke and touch, and look for people to interact with about this. Then i should be comfortable enough to start working on building a robocology, while the conceptual and philosophical framework undergoes continual refinement.
That's it. That shall be the rough direction i intend to take.

The most important thing i must do for myself is to maintain the faith in myself. Doubt is a highly effective destructive weapon.