Artificial Intelligence (AI): A series of notes

2005-06-20
“If it looks like a duck, and walks like a duck, and quacks like a duck, then it’s a duck” (Ancient wisdom)

Unless it’s a model of a duck.

Artificial Intelligence is model building – we want autonomous machines, but the best we can do is build models of autonomous machines.

Eg, an artificial ant could be made to behave like an ant in many ways, but not as an ant in an anthill, or capable of making more ants.

2015-10-21
It’s probably possible to make an artificial ant that behaves like an ant in anthill. We may even be able to make an artificial ant that can reproduce in some way.

However, “behave like an ant” is not well defined. There are too many behaviours, and some are obviously easier to mimic than others. Nevertheless, it will soon be possible to make an ant-size robot that can navigate like an ant, climb vertical surfaces like an ant, etc.

But it will always be a model of ant, and therefore its behaviour will in some respect will not be antlike, and in other respects will be a bad imitation of ant behaviour. That’s simply the nature of models. Models are mixtures of emulation and imitation.

2016-05-15
Intelligence is even less well-defined than “ant behaviour”. We can mimic some intelligent behaviours, eg, sorting, learning correlations, recognising patterns, and so on, which are useful to augment human tasks such as diagnosis of a fault or illness, or finding the data we want. If a task is well enough defined, we can build a machine to do it.

But that’s the problem: “Intelligence” is simply not well enough defined. My notion of it is the ability to apply and adapt existing knowledge and insight to unanticipated problems. Every term in that definition is fuzzy and vague. Anyhow, some people (including me) would argue it’s more of a definition of creativity than intelligence.

Is consciousness part of  “intelligence”? Many people would say it is. A machine that merely solves problems isn’t intelligent, it’s just an algorithm. It’s not enough to know how to do long division, you have to be able to recognise when and why you should do it. An intelligent entity then would be able to apply the rules of the algorithm to another problem. This claim entails that intelligence can abstract rules and patterns, and recognise them in different contexts.

“Understanding” is another component of intelligence. Isn’t it? Well, it does have something to do with learning: an intelligent person is one who can make sense of new explanations. “I don’t get it” at one extreme means “I haven’t figured it out yet”, at the other it means “I can’t figure it out”. The latter is a measure of intelligence.

And that’s just three attempts to make sense of “intelligence”. We’re long way from knowing exactly what we mean by “artificial intelligence”. Far enough that we may not even recognise it when we see it.

The recent development of “deep learning” artificially intelligent neural nets crystallises the problem. It’s already clear that we can evaluate the results of their operations, but we can’t figure out how they do it. What’s more, they have come up with solutions that humans have not only not produced, but have trouble recognising as viable solutions. For example, some AIs are better then humans at recognising cancerous tissue.

2018-07-03
If we accept “intelligence” as a label for problem-solving abilities, then consciousness is not required. That makes the neural-net AIs more than a little spooky.

Accidentals and Essentials: Experience and Reality

Observation and Theory, Experience and

Imagination: What’s Real and What Isn’t

1.0
We experience things and processes, events and spaces, times and moments, extended sequences of events, and so on. We may name (or label) any such experience, and having done so, we tend to think of it as a unified experience, whatever its extent in space and time. Naming is the first step to theory, the precondition of explanation.

2.0
2.1 The things we experience are bundles of sensory inputs. They may be relatively simple bundles, such as the ones that we label “apple”; or complex, hierarchically layered bundles, such as “song”; or more complex networks of experiences such as “fair play”; and so on. But all experiences which we perceive or apprehend as having some kind of unity in space and time can be reduced to a bundle of sensations.

2.2 The “unity in space and time” seems to be a given. Neuro-psychological research indicates that it’s a product of the brain’s processing of sensory inputs. Comparing human and other animals’ perceptions clarifies that insight: a frog responds to a fly-sized object such as a raisin only when it moves. But when it moves, the frog will try to catch and eat it, even when it’s a raisin.

3.0
3.1 We learn in grammar class that nouns name essentials (apple, kitten, rainstorm, triangle) while adjectives name accidentals (red, soft, greasy, warm, loud, sharp, salty, etc). The trick seems to be to recognise what makes an apple an apple and not a kitten; and what features of an apple can vary without destroying its appleness. But these distinctions are illusory, since any description of the essence of an apple is merely a list of accidentals, its so-called properties. Botanical classification makes that quite clear. We may infer that “apple” is a the minimal collection of accidentals that differentiates it from pear, cherry, raspberry, peach, mango, .... Note that the properties overlap: it’s the differences between the lists that differentiate the fruits.

3.2 We can also differentiate objects by abstract qualities, such animate/inanimate, food/non-food, etc. Anthropologists have discovered that there are no universal classification systems. While there is little doubt that we perceive objects as more or less stable collections of sensations, that does not entail that we classify them that way. Apparently humans have a penchant for believing that whatever they name is real. So when we name some aspect of human behaviour as “just”, we start collecting the properties of Justice, and arguing with other people about which of these aspects are accidental or essential. Thus the white wig and black robe of a British judge are accidentals. But is the British Common Law a better system for arriving at just judgments than the European Roman Law? Any argument pro or con will rely on explicit and implict assumptions about the essence of Justice.
 

4.0
4.1 The brain combines sensory inputs into experiences. Some of these combinations are built-in, so much so that the brain organises different sensory inputs into the correct temporal sequence even though the processing time of the inputs varies so much that the results are in not in the correct sequence. When the brain fails to produce the correct image of the reality mediated by senses, its possessor is more or less deluded. If he knows it, he may well be more disturbed by that knowledge than by the failure to parse reality correctly.

4.2 Some of these combinations are inherently incorrect: we call them “illusions”, and the essential (;-)) point is that knowing we perceive an illusion does not cancel it. Nor does the fact that we have acquired  many, perhaps most, of these illusory parsings during the development of the visual cortex after birth.

A Comment on Ayn Rand

(A repost with some amendments)

 Ayn Rand and her followers worship money. But her notions on money are such a muddled mix of insight and delusion that it's hard to know where to begin a rational critique. From the Ayn Rand lexicon (http://aynrandlexicon.com/lexicon/money.html):

     Money is the tool of men who have reached a high level of productivity and a long-range control over their lives. Money is not merely a tool of exchange: much more importantly, it is a tool of saving, which permits delayed consumption and buys time for future production. To fulfill this requirement, money has to be some material commodity which is imperishable, rare, homogeneous, easily stored, not subject to wide fluctuations of value, and always in demand among those you trade with. This leads you to the decision to use gold as money. Gold money is a tangible value in itself and a token of wealth actually produced. When you accept a gold coin in payment for your goods, you actually deliver the goods to the buyer; the transaction is as safe as simple barter. When you store your savings in the form of gold coins, they represent the goods which you have actually produced and which have gone to buy time for other producers, who will keep the productive process going, so that you’ll be able to trade your coins for goods any time you wish.

    As soon as she leaves the standard definitions of money (means of exchange, store of value/savings), her key points drift into nonsense. One of these misleading notions is that gold has “tangible value”. Gold has no intrinsic value; its value as money is only and exactly what people believe it is. The Conquistadores could not understand how the value of gold among the Inca and other South American peoples could be so low. They looted the gold, took it home to Spain, and promptly caused ruinous inflation. It was only when Spain used its gold for external trading that it could be used as capital. So much for the intrinsic value of gold.

     The notion that money somehow buys time for future production misses the point. "Delayed consumption" is possible only when there is a surplus of goods or productive capacity. Money cannot create a surplus, nor is it needed to ensure that any surplus will be used. Humans have invented many ways of saving surpluses without money. What’s needed to create a surplus is a technology that multiplies the effect of human work, such as agriculture. What’s needed to delay its consumption is a system of values and customs that will ensure the surplus (such as grain) will be stored for later use and trade. Neither of these require money.

     Fact is, even today much trade is done without money. The basic rule of all trade is "mutual obligation". The members of families and social circles trade goods and services because sharing is one of the obligations of these social groups. They keep pretty close track of their trades, too. Exact matches aren’t required, but everyone is expected to share their goods and provide services as best they can.

     And like practically everybody, Rand misquotes St. Paul’s comment on money. Later in the article, she says, So you think that money is the root of all evil? . . .  Have you ever asked what is the root of money?

     St. Paul actually wrote, The love of money is the root of all evil. Look it up!

    Money is a way of making trade with strangers possible, and thereby creating mutual dependence. A very useful invention. Eg, just try to calculate how many people have been involved in producing a 98 cent ball point pen and making it available to you. It’s made of several kinds of plastics and metals, which had to be mined, refined, processed, and shaped. The pen had to be packaged, warehoused, transported, and shelved. Even in a small town where you know most people, you aren’t necessarily a close friend of the shop owners and staff, but they serve you all the same.

     A stranger is someone to whom you owe nothing, and vice versa. This makes interaction between strangers dangerous. Hence, all societies have had to invent ways of making at least temporary mutual obligation possible. Think of "guest right", for example. You not only have the right to stay among your hosts, they have an obligation to protect you. A pretty good deal; you’d better have some good stuff to trade with them.

     So why do all those strangers work to produce and deliver a cheap pen to you? Because money makes it not only possible to trade with people you will never see, it makes it easy to do so.

     Nowadays, money trades are used to measure economic activity. They’re totalled in the Gross Domestic Product, or GDP, a number of such incomplete, gappy data that it causes pernicious delusions. The worst of these is the belief that an ever-rising GDP means we’re getting richer. Yet every time there’s major storm destruction, there’s a spike in GDP in the following months and years as the damage is repaired. I don’t think having to repair storm damage is making us richer. Besides, even in our highly monetised economy, about 1/3rd of economic activity does not involve money. In pre-money times, that was 100%.

     There is one value to the GDP: it can tell you how much of your spending eventually ends up in various pockets. Eg, in Canada, we spend about 10% of our GDP on healthcare. Every time we spend a dollar, a dime wends it way to the healthcare providers. Some of it gets there via taxes, some via insurance payouts, some via personal spending. So we could describe our GDP in terms of these three types of spending, but then we wouldn’t know what exactly the taxes, insurance payouts, and personal spending bought. What the GDP means all depends on how you analyse it. And that means that people with different axes to grind will analyse it differently.

     Basic rule about money: Money and wealth flow in opposite directions. This should be obvious, but most people tend to think that more money means more wealth. Money is only potential wealth, a point Rand doesn’t get quite right either, although her notion of money as buying time for future production comes close. But anyone who’s lived through extreme inflation, or has absorbed the stories of the people who did, knows that money isn’t wealth.

     I think everybody needs a good introductory survey course in anthropology, to learn about all the ways humans have organised the production and distribution of goods and services. It might cure one of the notion that our current economic arrangements are somehow natural or god-given. For over 95% of our existence as a culture-creating species, we humans have had no money. Yet humans managed to produce the goods and provide the services they needed. It’s true that money, because it accelerated trade, and more importantly enabled trade with strangers, accelerated the creation of wealth. But trade, and its beneficial effects on wealth creation and cultural exchange, existed long before money.

     2013-03-11 / 2014-05-23 / 2015-11-07 / 2018-07-03 / 2019-10-15

A Grammar checker.

I’ve just tried Grammatik, a grammar checker built into WordPerfect. To say it’s bad is an understatement. The only actual error it flagged was a typo. All the other errors showed that Grammatik could not parse sentences correctly, and so mistook verbs for nouns and nouns for verbs; demanded modifiers where none were needed; confused proper names with common nouns; didn’t recognise objects; and on and on and on.

One example: “A voice from on high spoke:...” Grammatik said it should read “A voice from a high spoke...”

I don’t know if the other wordprocessors have equally bad grammar checkers. I suspect so. Avoid them, they will thoroughly mislead you.

BTW, Grammatik flagged eight errors in this short screed.

Oliver Sacks: The River of Consciousness

     Oliver Sacks. The River of  Consciousness (2017) Posthumous collection of essays assembled from notes and edited from drafts. As the title suggests, Sacks is thinking about consciousness, the hard problem of philosophy and neurology. In one essay he wonders if consciousness is a discontinuous sequence of moments.
    Vision seems to be discontinuous he says, citing experiments measuring the response times to changes in the visual field. I think it's obvious that vision is discontinuous: the light-sensitiva molecules in the retinal cells decay when struck by light, and must be rebuilt in order to decay again. This process takes about 1/10th of a second. Then the nerve signals generated must be processed so that the human can see. This can take longer than a 1/10th of a second, since objects must be recognised, etc. The fastest conscious reaction times to an expected visual stimulus is about 1/6th of a second in children and teenagers, and double or triple that in adults. Responses to unexpected visual stimuli take much longer. Thus the visual contents of consciousness are constructed from perceptions that take a sizeable time to assemble. the same applies to the other senses. The unbroken stream of consciousness is an illusion. It can be a dangerous one, since what feels like an instantaneous reaction takes at least half a second. At highway speeds a car travels about 14 metres in 1/2 a second.
     Sacks also has interesting observations of the subjective passage of time. His patients vary enormously in the rate at which they process sensory information, and that processing relates to the feeling of time passing. He tells of taking photographs several  minutes apart of one of his post-encephalitic patients, then binding the prints into a flip-book, and seeing the patient slowly lift his arm. Conversely, some of his patients entered a high-speed phase, and reported that they found the world around them moving unbearably slowly.
     The book feels unfinished. Most of the essays consist of extended notes. Sacks didn’t have time to rewrite for continuity, style and clarity, and this sometimes shows in a banal or cliche phrase. However, for any fan this is an essential book. For the general reader it serves as a very good introduction to some of the conundrums of consciousness and mind. ***½
     Correction 2018-11-10: At highway speeds of about 100km/h, a car travels about 14 metres in 1/2 a second.

If you want ot be a writer: Stephen King On Writing

     Stephen King. On Writing (2000). Subtitled “A Memoir of the Craft”. I’m not a fan of Stephen King, not because he’s a bad writer, but because horror fantasy doesn’t move me. I read some of his short stories way back when, and thought they were well done.
     King writes both about the nuts’n’bolts (grammar and style, narrative pace, character, etc) and his own experience as a writer. The book is worth reading for both. If you need some guidance to improve your writing and work habits, read this book. If you need some inspiration and emotional support because you’re not sure you can hack the writing life, read this book. You will improve your mastery of the craft, and you may discover your writing groove. Or you may discover that you’re not a writer after all. Either way, the book is worth reading.
     More take-aways: Writing is a compulsion, it’s what you have to do to maintain your sense of self.
     Reading a lot is essential to your development as a writer.
     A story is out there, like a fossil to be discovered. Writing it is uncovering the fossil.
     Interesting for any Stephen King fan, and for anyone who's curious about the writing life. ****

What did your lfe mean? The Five People You Meet in Heaven (Mitch Albom)

     Mitch Albom. The Five People You Meet in Heaven (2003) Albom made his name with Tuesdays With Morrie, which was made into a successful movie. I’ve seen the movie, it teeters just this side of sentimentality.
     This book (also made into a movie) teeters over, which is a pity, since it’s a lovely idea: Eddie, the hero, dies while saving a little girl when a gondola on an amusement park ride falls. The story tells of how in the afterlife he meets five people who affected his life in ways he didn’t fully understand or didn’t know. He needs to discover how his life made sense and had a purpose before he can live in his own corner of heaven.
     Eddie had a harsh upbringing, went to war, came home a changed man, and didn’t have the children he and his wife wanted. He ends up working in maintenance in the Ruby Pier Amusement Park, a job his father held, and which he thinks marks him as a failure. The five people he meets show him otherwise.
     Albom writes well, if occasionally too consciously ironic, and with sometimes too much authorial commentary. If the book causes the reader to reflect on how minor and major incidents shaped his own life, it will have succeeded. As a story about a likeable man who finally understands his own value, it’s well-done. Read it. ***