On Intelligence
Ever since I read "On Intelligence", a must-read book from Jeff Hawkins, I feel like I can find an explanation for everything that goes on in our minds. So this page is just a compendium of random thoughts about our psyche. Hawkins' theory of intelligence is quite simple, but precisely because of this, very powerful. I read the book to learn new things about Artificial Intelligence, but I found out that understanding your brain is actually possible. For those of you who haven't read the book, I will make a small synthesis here. Hawkins defines Intelligence using what he calls the Memory-Prediction framework. While traditionally, intelligence has been defined in terms of intelligent behaviour, he assumes that this is wrong, and that the Turing test of intelligence is pointless. Even if you are fooled by a machine, it doesn't mean that the machine is intelligent. The brain, or precisely, the neocortex, is not a CPU that gets some inputs, process them, and generates some outputs. The neocortex, where intelligence resides, is a memory system, but a very special one. The neocortex has more connections going backwards than those going forward. These connections have been usually ignored, but Hawkins gives a plausible meaning to them: the ability to make predictions. Our brain is just a memory that makes predictions and tries to fit the input with the predicted output. So when the input comes, we can almost immediately know if that is what we expected or not. Faster silicon chips performs worse than neurons because they require millions of instructions to do the same. Of course, the input modifies the neurons configurations, since our memory works with invariant representations. This is the basic idea of this theory. If you think a bit, you will be marvelled that everything you have learned matches perfectly in this theory. In the book you can find some nice examples to illustrate this theory. What I will do here is what I have found myself as plausible explanations for psychological and philosophical matters. Far beyond my interest in Computer Science... Notice that even if I don't write "I think" all the time, what I write here is by no means founded in any solid base. It's just deductions I have made myself using the memory-prediction framework as a base.
|
||||
On Psychology Before my sister gets angry, I must say I know nothing about Psychology. What I will try to explain here are some possible answers to "why we do what we do", using the theory mentioned above, rather than explaining things about Psychology. I just needed a catchy phrase . When we are born, our neocortex has a structure, but it's basically empty, waiting for information. We learn, and our invariant representations, or so to speak, our model of the world is formed. This model will be modified and extended while we grow. But as our brain gets "full", new experiences will be accomodated to match known ones. A new movie will remind us of some mixture of old ones, and we will soon forget the details of this new movie, while a child will remember it vividly. So if you want to know the explanations of your fears, a therapist will start digging into your past to know that world you started creating when you were a child. Regression will help you to understand why do you do what you do now. Hawkins points out that our feelings are not part of our intelligence. An intelligent machine doesn't need to feel anything. Feelings are part of the primitive brain, and they are just like some kind of chemical factories. They are an input to your brain and their output is your blood. So, as any other experience, we will form invariant representations of our feelings. What does it mean to love, to hate, to be happy, to be sad...? Each person may perceive these emotions in different ways, and act differently to them. How prediction works in this case? Very easy and neat, indeed: your mind predicts how you are going to feel. In the prediction framework, if a prediction is not matched, it will alert our visual system or whatever the sensation that originated the input. For feelings, I think that it will alert our "chemical factories", and eventually trigger an opposing reaction. For instance, imagine that I cook a cake for you every day. You are happy with that. But today, you were waiting for the cake, and I forgot to make it. Your brain was expecting not only the cake, but also expecting "to be happy". So it will trigger another feeling, like being sad or angry. The feeling triggered also will depend on your past experiences. Notice that you don't specially feel "happy" everyday because of the cake. It just became the normal state. But now imagine the opposite cake, sorry, case . You haven't eaten a hand-made cake for centuries. And today, I just decided baking one for you. Your brain don't expect it, so it will alert your factories and you'll probably be a bit happier today . Now, you get it. Children are easily surprised because their minds are not so full of memories, so they can not make as many predictions as we do. You can make them happy very easily, and they can get angry so easily too! Now, you got me. Why I don't look moved by your birthday present? Why I didn't get scared in that movie? Why I looked so bored yesterday? Predictions. Does it mean that new things will make us happier? Not necessarily. As I said, being happy or not it's also a part of the model of the world in each person. When we fail our predictions, our senses have to be alert. We predicted to climb that step, but our foot didn't feel it under its sole. Be careful! You can not stay calm. Too much of this will produce anxiety. While young people likes to experience new things, usually old people just want to feel secure by staying at places they know. It's hard for me to plan trips. If we talk about going to a new place, or doing something I don't know how to do, like skiing, I'll start to feel anxious about it. Probably I'll just think I better stay home. But then, if you tell me to go to a place I've already been to, I may say that's boring and that I don't want to expend money for experiencing the same things again. More than that, when we see or feel something we have never seen or felt before, the first thing our brain will do is trying to match it with some invariant representations. We don't see a moving and noisy signal, but the result of our prediction plus the signal. In a sense, we try to put order to the chaos. We see the world, so that we can cope with it. Is this love, or just some friendly feeling? Probably we will need a closer look and more experiences to understand the difference and modify our model. If you understood this, then you can understand better why people do what they do, or why they feel like they feel. Starting from you. Well, maybe you are reading this to learn something about me. In such case, I'll left some open thoughts here. I think I expect too much from people, including me. I'll be frustrated if I can't solve things as I predicted them to be perfect. Why? Maybe as a child I was always told that I'm gonna be the best, but that may just have helped to grow my insecureness. What about others? Well, I try to be nice to people, because if I know that people around me are not happy, then I can not be happy. But at the same time, of course, I my brain is making a lot of predictions, based on the way I would act, or the way my family or close friends would act. If I stop doing the washing up because you ask me to take a look at your computer, and then I finish and have to leave in a hurry, I probably will expect that you finish the washing for me. If you didn't, I won't be angry, but I'll be disappointed. Of course, you won't know, because I don't want to tell you, because I just want you to be happy, or I'll feel sad if I see you upset or angry. Now, you can call me obsessive-compulsive, of course . The question that arises now is if it does us any good to understand how our brain works. Sometimes we are happy being able to formulate questions, but feel desolated when we find the answers. Ok, you found out the quid of your traumas, but how to overcome them? Going to psychotherapy, for instance. I found out that I am obsessive because I predict too much. What should I do? Maybe prevent my mind from doing so? How? Well, perhaps changing behaviour will help, since behaviour it's an act of intelligence. By changing the output, we may be able to alter the predictions. Make the unpredictable happen. You may be upset, or maybe not, but I better complain about you not doing the washing up. At least others will know what do I expect from them. Also, I can try to distract my mind with other activities, like writing this page. Unfortunately, I really do think that ignorance is the key to happiness. Make your own conclusions, and help me out . Here's a nice Japanese quote from School Rumble...
|
||||||
On Oneirology While discussing with a friend I was told that this theory is not very sound because while dreaming we do not get surprised if a familiar room has changed. Well, I think it proves the opposite. So let us talk about dreams a bit. The theory explained in "On Intelligence" does not deal with the subconscious, although I think it points out about the origin of Imagination. But since the theory explains the global functionality of the brain, it should be easy to explain every example you can think of. In the case of imagination, it can be seen as the process of building new invariant representations from older invariant representations or memories. It's playing with the images and ideas in our brain. We do this consciously. What about the subconscious? What happens when we are dreaming? I think that the answer is very simple with this model (that's why is so powerful! ); we are playing again with invariant representations. The brain never stops, so even when we sleep, it wants to do something. But then, we do not receive much sensorial input, so all the inputs are from inside our mind. Basically, we imagine things, playing with known objects, situations, or ideas, and use them as input. Now let's go back to my friend's room. He imagined a completely different room from the one he knows in reality, but he is not surprised in the dream about that fact. Of course not, since that room was the one he predicted in his dream, or better say, the prediction system made up an imaginary room. The output matches the input, so there is nothing to be surprised of. I know what you are going to say next: "but I DO get surprised in my dreams". Of course the process of imagination leads us naturally to the unexpected. That's what it is there for. When we are trying to solve a problem and finally find a method to solve it, we are both pleased and surprised. Maybe surprised is not the right word, but if you read the previous topics, you will know that by "surprise" I mean "unmatched prediction". It sufficiently clear to me: we play with our knowledge, and eventually we "deduce" a new idea, we "imagine" a new picture, or we "dream" new situations; and then we write it down (like I am doing right now ), we draw it on paper, or we tell it to friends. If you check the Wikipedia, it says: "It would appear that mammals are the only, or at least most frequent, dreamers in nature, which is perhaps related to their sleep patterns". That also makes a lot of sense to me, since mammals are the only animals with neocortex, the origin of intelligence. As a final word, I should say that I do not usually remember dreams myself. That's probably because I do rest well . So, maybe you do not agree in reducing dreams to just a "subconscious imagination". Contact me for further discussion . Further readings:
|
||||||