13 April 2010

Sense Making

Whether it makes sense to them or not, all living organisms from bacteria to fishes to plants to animals sense their environment one way or other. Over the course of evolution, animals developed a complex sensory organs to sense their environment, and brain to make 'some' sense of it and react accordingly. The brain is fundamentally an information (signal) processing system like a computer but with different architecture.

The brain circuitry (neural network) is achieved in two ways. Evolutionarily low level animals have innate hard-coded brain circuitry. Each part of its circuitry is achieved by complex set of genes. As it is hard-coded, their behavior is highly inflexible and as each part of it requires complex set of genes, it can only accommodate relatively simple and straight forward behavior. Apart from this innate hard-coded brain circuitry, evolutionarily high level animals have also developed soft-coded brain circuitry. The soft-coded brain circuitry can be almost blank at birth, but requires some mechanism for learning to create appropriate circuits (connections). As it is blank at birth, the genes are mostly needed for learning mechanism only. It is flexible and new behaviors and skills can be learned. But it takes time to learn and requires long offspring developmental period. And hence essential parental-care is co-evolved! Compare to a deer offspring that can start walk and run within few hours after birth, a human child requires the longest developmental period.

As it is easily observable, we notice how children learn to walk and speak. But we also learn to see, hear, feel, smell and taste the same way. As we learn, we build a complex hierarchy of models in our brain to make 'some' sense of the world we live in. This kind of learning is highly innate and automatic. This is how most animals (mammals and primates) learn. Apart from this kind of automatic learning, humans have also developed a capacity to learn to learn – i.e. not just learn, but how to learn, how to learn efficiently, systematically and flexibly – to learn deliberately. This allowed us to look at the world objectively.

Though our brain has the capacity to learn deliberately, it comes with so many limitations. Our brain can only keep 4 to 7 items in its short-term or working memory at a time (monkeys and even pocket calculators exceed this capacity. I wish, I can remember a phone number easily for a minute!). Our brain can only focus (consciously) on one thing at a time. Though our automatic learning uses fantastic statistical algorithm, in our deliberate learning side we are so bad at even simple statistics. We cannot intuitively understand if it is too big or too small; electron mass (9.10938 * 10-31 kg) and proton mass (1.67262 * 10-27 kg) are just too small to us to make any relative comparison, likewise few millions and billions are just too big to us. It is very hard for us to make good sense of vast cosmological and geological time. Hence we find it difficult to appreciate and intuitively understand evolution. Though our brain has the capacity to learn objectively about the world, it is not natural or straight forward to us. It requires systematic approaches, many tools and long sustained efforts to really make sense!

However marvelous our brain might be, with its limited capacity, we have to divide things into many meaningful levels for understanding. At each level we have to analyze individual pieces (reductionism) and look at how individual pieces combine together as a whole (holism) where sometimes totally new patterns, behaviors and properties emerge like a magic (emergence). Watch a cute child laughing on a TV screen. This emerges from a collection of dots or pixels each with different colors and brightness that varies with time. Likewise our mind emerges from a huge collection of brain cells each with simple operations. If you want to appreciate and enjoy the fantasy world of Harry Potter book, it requires understanding at many levels: first understand the alphabets, then how collection of alphabets forms words, then how collection of words forms sentences, then paragraphs, and then chapters. Likewise it requires many meaningful levels to really understand any complex thing such as human mind or the world we live in. The elementary particles give rise to atoms and forces; the atoms combine and form molecules and some molecules combine and form cells; a certain collection of cells create cognition, a mind! That is, the world including galaxies, stars, planets, life, mind, etc. is emerged from vast collections of simple things. But it is not easy for our mind comprehend the whole thing. It requires a systematic understanding at many levels such as physics, chemistry, biology and psychology. Though we have many gaps in our understanding at present, it is clear to see this from nature’s symmetry at many levels; and nature's re-confirmation about her nature again and again at many levels. But just sensing, feeling or believing is not enough to understand her; it requires a systematic deliberate learning.


Anonymous said...

The brain circuitry (neural network) is actually achieved in more than two ways - the innate hard-coded, soft-coded, and intelligence. Intelligent brain cells are responsible to deduce data provided by the other two – the hard-coded and the soft-coded parts. The process is also known as independent thinking as intelligent brain cells are responsible for analyzing information to make independent choice(s). We still know very little about how exactly that part works at neural level.

RajK said...

Intelligence is an emergence behavior and it requires understanding at many systematic levels; though, its low level (neuronal or cell level) is reasonably understood, the complexity and gaps in our understanding exists in the middle levels. But intelligence is all about inferring from partial information. To make any reasonable sense, as I mentioned in the article, just sensing, feeling or believing is not enough, a systematic deliberate learning is required. But that is not natural or straight forward. Thanks for proving it!