This Much I Know (about CS and AI)

Inspired by The Guardian column of the same name, I decided to list here my key learnings of the last several years regarding Computer Science and Artificial Intelligence (AI). Few of these are my own insights, and I welcome comments and responses. From arguments I have had, I know that some of these statements are controversial; this fact surprises me, since most of them seem obvious to me. Statements are listed, approximately, from the more general to the more specific.

Catwoman, my old flame

Those of you paying attention to these lectures will realize how obsessed I am with Economics.  That flaxen-haired lady promised so much, but she has so many flaws and failings.   When we first meet her, it seems she is everything you could wish for:  she is concerned with how society should be organized, how people should be given material goods, how the benefits of new technology and material well-being should be shared with all, and how the poor should be enriched, so that they can spend their time on self-improving and fulfilling activities, like art and sport.  So much is promised!
But then, once the flirtation and seduction are over, her flaws become evident. I have been thinking about these flaws again, having just read Deirdre McCloskey’s superb 2002 pamphlet, The Secret Sins of Economics.  Many of McCloskey’s criticisms are ones I (and many others) have made before, but some are new.   I decided, for comparison, to list here my chief complaints with this blemished beauty, this feline seductress, Our Lady of the Catallacts.  Date her if you wish, but you should read these accounts by her ex-lovers before you do.
First, she is blinkered, often unable to see what is obvious to anyone else – that we are all shaped by social and cultural forces, and peer pressures.   Instead, Catwoman and her acolytes invariably assume an individualist explanation for any economic or social phenomenon, and then seek to demonstrate it.  McCloskey calls this a focus on the P-variables (price, individual prudence, profit, the profane) as distinct from the S-variables (solidarity, speech, stories, shame) which Anthropology, that Indiana Jones of academic disciplines – creative, unruly, a thorn in everyone else’s side – has focused on.   A classic example is Levitt and Dubner’s Freakonomics.
Because of her blindness to the social, Cat Lady mostly ignored (until recently) major aspects of society, such as Institutions, legal frameworks, norms, and power relationships, aspects which can make or fail the marketplaces she says she studies.   She can’t claim that no one mentioned these to her, since 19th-century economists such as Karl Marx made the study of these aspects the work of a lifetime, and their study has continued to the present by sociologists and anthropologists and political scientists.
She has also been blind to anything historical or temporal, as if all her work stood outside the mundane and messy world in which we live.  This blindness manifests itself most strongly in the complete disregard (until recently) for endowments:  how did we get to where we are?  So, for example, free trade theory says that if England produces textiles more cheaply than Portugal, and Portugal produces wine more cheaply then England, the two should trade textiles for wine, and wine for textiles.   And the choice of these products is a subtly clever one, obfuscating much, since wine needs sunshine and not too much rain, while textiles (in the 18th and early 19th centuries) needed lots of rain, in order that the damp air would ensure cotton threads did not break when woven by machines.   So, Portugal’s sunshine and Northern England’s rain, being part of the God-given climate, were natural advantages, beyond the control or manipulation of any temporal human powers.  Free trade seems to have been ordained by the Almighty. But why consider only England’s textiles and not Ireland’s?    The answer is that Ireland had no textile industry to speak of.  And just why is that?  After all, much of Ireland is as damp as the valleys of Lancashire.   The reason is that the owners of northern English textile factories lobbied the British authorities to exclude Irish-made textiles from entering England.  When Ireland lost its own Parliament in a hostile takeover by Westminster, this protectionism for English textiles was entrenched, and the growing British Empire provided the critical masses of customers to ensure bonuses in Bury and Bolton and Burnley.     (Is it any wonder that people in Ireland and India and elsewhere sought Independence, when colonialism so powerfully stifled economic aspirations.)  Northern England has no natural comparative advantage in textile production, at least, not when compared to Ireland, but an artificial, man-made advantage.  The same type of advantage, in fact, that South Korea today has in ship-building, or the USA in most computer and aerospace technologies.   Where, in the mainstream theory of free trade, are these aspects studied, or even mentioned?
And when, angered by these failings, you face her with them, the wench promises you that that was all in the past, and she will be different from now on.  Path dependence and network goods and institutional economics are all the rage, she says.   But then you find, she’s still up to her old tricks:  She says she’s building models of economic phenomena in order to understand, predict and control, just like physicists do.  But, although it looks like that’s what she’s doing, in fact her models are not models of real phenomena, but models of stylized abstractions of phenomena.  Her acolytes even use that very word – stylized – to describe the “facts” which they use to calibrate or test their models.
Of course, she will say, physicists do this too.  Newton famously assumed the planets were perfect spheres in order to predict their relative movements using his theory of gravitation.   But physicists later relax their assumptions, in order to build revised models, in a process that has continued since Newton to the present day.  Physicists also allow their models to be falsified by the data they collect, even when that data too is stylized, and overturned.     Instead, Catwoman is still assuming that people are maximizers of individual utility, with perfect foresight and unlimited processing capabilities, obeying the axiom of the irrelevance of independent alternatives, when all these assumptions have been shown to be false about us.   When was the last time a mainstream economic model was overturned?
Indeed, here is another of her flaws:  her loose grasp of reality.  She says we are always, all of us, acting in our own self-interest.  When you quiz this, pointing out (say) a friend who donated money to a charity, she replies that he is making himself feel better by doing something he thinks virtuous, and thus is maximizing his own self-interest.  Her assumption, it turns out, is unfalsifiable.   It is also naive and morally repugnant – and false!  Anyone with any experience of the world sees through this assumption straight away, which is why I think our feline friend is borderline autistic.   She just does not know much about real people and how they interact and live in the word. Who would want to step out with someone having such views, and unable to reconstruct them in the light of experience?
And, despite her claims to be grounded in the material world (Paul Samuelson:  “Economics is the study of how people and society end up choosing, with or without the use of money, to employ scarce productive resources that would have alternative uses,  . . .”), she sure is fond of metaphysical entities for which no hard evidence exists:  invisible hands, equilibria, perfect competition, free trade, commodities, in fact, the whole shebang.   As marketers say, the existence of a true commodity is evidence that a marketing manager is not doing his or her job.  In comparison, Richard Dawkins with his memes is a mere amateur in this creation of imaginary objects for religious veneration.
One could perhaps accept the scented candles and the imaginary friends if she was a little more humble and tolerant of the opinions of others.  But no, the feline femme fatale and her acolytes are among the most arrogant and condescending of any academic disciplines.  Read the recovering Chicago economist McCloskey for an account of this, if you don’t believe me.   McCloskey’s anecdotes and experiences were very familiar to me, especially that sneer from an economist who thinks you’ve not acted in your own self-interest – for example, by helping your colleagues or employer with something you are not legally required to do.  Indeed, the theft by economists from philosophers of the word “rational” to describe a very particular, narrow, autistic behavior is the best example of this.   Anyone whose behavior does not fit the models of mainstream economics can be thus be labeled irrational, and dismissed from further consideration as if insane.
Date her at your peril!  You have been warned!

Metrosexual competition

Writing about the macho world of pure mathematics (at least, in my experience, in analysis and group theory, less so in category theory and number theory, for example), led me to think that some academic disciplines seem hyper-competitive:  physics, philosophy, and mainstream economics come to mind.  A problem for economics is that the domain of the discipline includes the study of competition, and the macho, hyper-competitive nature of academic economists has led them, I believe, astray in their thinking about the marketplace competition they claim to be studying.  They have assumed that their own nasty, bullying, dog-eat-dog world is a good model for the world of business.

If business were truly the self-interested, take-no-prisoners world of competition described in economics textbooks and assumed in mainstream economics, our lives would all be very different.  Fortunately, our world is mostly not like this.   One example is in telecommunications where companies compete and collaborate with each other at the same time, and often through the same business units.  For instance, British Telecommunications and Vodafone are competitors (both directly in the same product categories and indirectly through partial substitutes such as fixed and mobile services), and collaborators, through the legally-required and commercially-sensible inter-connections of their respective networks.  Indeed, for many years, each company was the other company’s largest customer, since the inter-connection of their networks means each company completes calls that originate on the other’s network; thus each company receives payments from the other. 

Do you seek to drive your main competitor out of business when that competitor is also your largest customer?   Would you do this, as stupid as it seems, knowing that your competitor could retaliate (perhaps pre-emptively!) by disconnecting your network or reducing the quality of your calls that interconnect?  No rational business manager would do this, although perhaps an economist might.

Nor would you destroy your competitors when you and they are sharing physical infrastructure  – co-locating switches in each other’s buildings, for example, or sharing rural cellular base stations, both of which are common in telecommunications.   And, to complicate matters, large corporate customers of telecommunications companies increasingly want direct access to the telco’s own switches, leading to very porous boundaries between companies and their suppliers.   Doctrines of nuclear warfare, such as mutually-assured destruction or iterated prisoners’ dilemma, are better models for this marketplace than the mainstream one-shot utility-maximizing models, in my opinion.

You might protest that telecommunications is a special case, since the product is a networked good – that is, one where a customer’s utility from a particular service may depend on the numbers of other customers also using the service.    However, even for non-networked goods, the fact that business usually involves repeated interactions with the same group of people (and is decidely not a one-shot interaction) leads to more co-operation than is found in an economist’s philosophy.  

The empirical studies of hedge funds undertaken by sociologist Donald MacKenzie, for example, showed the great extent to which hedge fund managers rely in their investment decisions on information they receive from their competitors.  Because everyone hopes to come to work tomorrow and the day after, as well as today, there are strong incentives on people not to  mis-use these networks through, for instance, disseminating false or explicitly-self-serving information.

It’s a dog-help-dog world out there!

Reference:
Iain Hardie and Donald MacKenzie [2007]:  Assembling an economic actor: the agencement of a hedge fund. The Sociological Review, 55 (1): 57-80.

Vale: Stephen Toulmin

The Anglo-American philosopher, Stephen Toulmin, has just died, aged 87.   One of the areas to which he made major contributions was argumentation, the theory of argument, and his work found and finds application not only in philosophy but in computer science.
For instance, under the direction of John Fox, the Advanced Computation Laboratory at Europe’s largest medical research charity, Cancer Research UK (formerly, the Imperial Cancer Research Fund) applied Toulmin’s model of argument in computer systems they built and deployed in the 1990s to handle conflicting arguments in some domain.  An example was a system for advising medical practitioners with the arguments for and against prescribing a particular drug to a patient with a particular medical history and disease presentation.  One company commercializing these ideas in medicine is Infermed.    Other applications include the automated prediction of chemical properties such as toxicity (see for example, the work of Lhasa Ltd), and dynamic optimization of extraction processes in mining.
S E Toulmin
For me, Toulmin’s most influential work was was his book Cosmopolis, which identified and deconstructed the main biases evident in contemporary western culture since the work of Descartes:

  • A bias for the written over the oral
  • A bias for the universal over the local
  • A bias for the general over the particular
  • A bias for the timeless over the timely.

Formal logic as a theory of human reasoning can be seen as example of these biases at work. In contrast, argumentation theory attempts to reclaim the theory of reasoning from formal logic with an approach able to deal with conflicts and gaps, and with special cases, and less subject to such biases.    Norm’s dispute with Larry Teabag is a recent example of resistance to the puritanical, Descartian desire to impose abstract formalisms onto practical reasoning quite contrary to local and particular sense.
Another instance of Descartian autism is the widespread deletion of economic history from graduate programs in economics and the associated privileging of deductive reasoning in abstract mathematical models over other forms of argument (eg, narrative accounts, laboratory and field experiments, field samples and surveys, computer simulation, etc) in economic theory.  One consequence of this autism is the Great Moral Failure of Macroeconomics in the Great World Recession of 2008-onwards.
References:
S. E. Toulmin [1958]:  The Uses of Argument.  Cambridge, UK: Cambridge University Press.
S. E. Toulmin [1990]: Cosmopolis:  The Hidden Agenda of Modernity.  Chicago, IL, USA: University of Chicago Press.

Gray on Akerlof and Shiller

Philosopher John Gray has a review in the LRB of Akerlof and Shiller’s new book on the errors of mainstream economics, a review which mentions the sadly-neglected economist George Shackle.  Shackle, unlike most academic economists, actually worked in industry and Government and had made investment decisions, and knew whereof he wrote.

If Akerlof and Shiller’s grip on the history of economic thought is shaky, they also fail to grasp why Keynes rejected the idea that markets are self-stabilising. Throughout Animal Spirits they portray him as reintegrating psychology with economic theory. No doubt this was one of Keynes’s goals, but it is not his most fundamental revision of economic orthodoxy. Among his other accomplishments he was the author of A Treatise on Probability (1921), in which he tried to develop a theory of ‘rational degrees of belief’. By his own account he failed, and in his canonical General Theory of Employment, Interest and Money (1936) he concluded that there was no way anyone could make forecasts. Future interest rates and prices, new inventions and the likelihood of a European war cannot be predicted: there is no ‘basis on which to form any calculable probability whatever. We simply do not know!’ For Keynes, markets are unstable less because they are driven by emotion than because the future is unknowable. To suggest that the source of market volatility is unreason is to imply that if people were fully rational markets could be stable. But even if people were affectless calculating machines they would still be ignorant of the future, and markets would still be volatile. The root cause of market instability is the insuperable limitation of human knowledge.
Continue reading ‘Gray on Akerlof and Shiller’

The websearch-industrial complex

I think it is now well-known that the creation of Internet was sponsored by the US Government, through its military research funding agencies, ARPA (later DARPA).   It is perhaps less well-known that Google arose from a $4.5 million research project sponsored also by the US Government, through the National Science Foundation.   Let no one say that the USA has an economic system involving “free” enterprise.

In the primordial ooze of Internet content several hundred million seconds ago (1993), fewer than 100 Web sites inhabited the planet. Early clans of information seekers hunted for data among the far larger populations of text-only Gopher sites and FTP file-sharing servers. This was the world in the years before Google.
Continue reading ‘The websearch-industrial complex’

Anti-Krugman on free trade

Without intending to, I again came across an essay by Paul Krugman that I had first read 3 years ago,  in which he seeks to understand why anyone could be opposed to the claims of mainstream economics in favour of free trade, particuarly anyone intelligent and informed.   Unfortunately, Krugman fails at his self-appointed task, because he resorts to arguments which are essentially ad hominem attacks on the opponents of free trade theories.

In the context of the assessment of alternative public policies I view ad hominem arguments as invalid and irrational, a sign that a proponent cannot provide substantive justifications for his or her claims.  (Such personalist attacks are not necessarily invalid or irrational in all contexts.)  I believe economists do themselves and their profession a great disservice by dismissing arguments against free trade on the grounds of ignorance or insufficient intelligence, or any of the other ad hominem arguments Krugman uses.

There is a perfectly respectable intellectual argument, not based on ignorance, against Ricardo’s argument. This is that his theory ignores the single most important question in the debate, which is: Where does a comparative advantage come from? A national comparative advantage is almost never, despite what many free trade advocates seem to think, an act of God. It is almost always the result of human actions, witting or unwitting. The Republic of Korea (South Korea) has no “natural” reason to be a world-leading steel-producer or world-class large-ship-maker. But successive ROK Governments intended the nation to lead these industries, and the nation has in consequence succeeded in doing so: through direct government investment; through subsidies to Korean companies; through discriminatary trade practices against foreign companies; and through large-scale, concerted, long-term public and private pressure on private Korean companies to invest in particular ways, and not in other ways.

I note in passing that a strong national desire to establish an international lead in a particular commercial sector does not necessarily mean it will happen. Japan’s powerful Ministry of International Trade and Industry (MITI), when it existed, tried repeatedly to establish Japan as a global leader in avionics, electronics used in aircraft, without any success at all.

So, imagine you are a political leader of a country with an existing large-ship-building industry faced with rising Korean competition, for example the UK in 1980. Your chief economic advisor tells you that free trade is an unmitigated good, and that your markets should all be open, because the economic benefits outweigh the costs. These economic benefits may outweigh the associated costs or they may not in any particular case, and they may well fall differentially upon different stakeholders with different political power.  Even if the economic benefits outweigh the economic costs, the social cost-benefit analysis may be very different (eg, long-term unemployment; increased crime as a result of unemployment; the destruction of local communities; the dispersal or loss of a skills base; the loss of defence-related industries and companies; etc).

Whatever the cost-benefit outcomes, what really sticks in my craw is that the advice from economists is to do nothing to protect local industrial capabilities, so that others who have ignored such advice (eg, the Government of the ROK) should prosper.  You would think that, of all people, economists would have learnt some game theory!
The argument here is not between the ignorant and the wise, as Krugman seems to think, but between people with conflicting sets of values. My values tell me that economic theory should include discussions of political power, discussions of social, cultural and institutional structures, and discussions of the reasons for and the consequences of differing preferences over outcomes; most mainstream economists do not think these issues are part of economic theory.

To be called ignorant because I have different set of values to an economist is not only an insult, and not only disrespectful of my rights as a stakeholder in the society in which I live; it is also revealing of the vapidity of mainstream economics, busy building abstract theories which ignore all the most important questions.
Economists have heard these criticisms of free trade theory since at least the time of Marx. Is it recalcitrance or malfeasance that allows mainstream economists to write articles as if ignorant of such criticisms?

In these times

You are a man of leisure, a sleepwalker, a mollusc.  The definitions vary according to the hour of the day, or the day or the week, but the meaning remains clear enough: you do not really feel cut out for living, for doing, for making; you want only to go on, to go on waiting; and to forget.
Such an outlook on life is generally not much appreciated in modern times:  all around you, all your life, you have seen the esteem in which action is held, and grand designs, and enthusiam:  man straining forward, man with his gaze fixed on the horizon, man looking straight ahead.  A clear gaze, a powerful chin, a confident swagger, stomach held in.  Staying power, initiative, strokes of brilliance, success:  all of these things map out the too transparent part of a too examplary existence, constitute the sacrosanct images of the struggle for life.  The white lies, the comforting illusions of all those who are running on the spot, sinking deeper into the mire, the lost illusions of the thousands left on society’s scrap heap, those who arrived too late, those who put their suitcase down on the pavement and sat on it to wipe their brow.  But you no longer need excuses, regrets, nostalgia.  You reject nothing, you refuse nothing. You have ceased going forward, but that is because you weren’t going forward anyway, you’re not setting off again, you have arrived, you can see no reason to go any further” (pages 142-143)

Reference:
Georges Perec [1967]:  A Man Asleep.  (Translation by Andrew Leak published 1990 in London, UK by The Harvill Press.)

Black swans of trespass

gould-blackswan
Nassim Taleb has an article in the FinTimes presenting ten principles he believes would reduce the occurrence of rare, catastrophic events (events he has taken to calling black swans).  Many of his principles are not actionable, and several are ill-advised.  Take, for instance, # 3:

3. People who were driving a school bus blindfolded (and crashed it) should never be given a new bus.

If this principle was applied, the bus would have no drivers at all.   All of us are driving blindfolded, with our only guide to the road ahead being what we can apprehend from the rear-view mirror.  Past performance, as they say, is no guide to the future direction of the road.
Or take #6:

6. Do not give children sticks of dynamite, even if they come with a warning.  Complex derivatives need to be banned because nobody understands them and few are rational enough to know it. Citizens must be protected from themselves, from bankers selling them “hedging” products, and from gullible regulators who listen to economic theorists.

Well, what precisely is “complex”?  Surely, Dr Taleb is not suggesting the banning of plain futures and options, as these serve a valuable function in our economy (enabling the parceling and trading of risk).  But even these are too complex for some people (such as those farmers, dentists, and local government officials currently with burnt fingers), and surely such people need protection from themselves much more so than the quant-jocks and their masters on Wall Street.  So, where would one draw the line between allowed derivative and disallowed?
Once again, it appears there has been a mis-understanding of the cause of the recent problems.  It is not complex derivatives per se that are the problem, but the fact that many of these financial instruments have, unusually, been highly-correlated.  Thus, the failure of one instrument (and subsequently, one bank) brings down all the others with it — there is a systemic risk as well as a participant risk involved in their use.   Dr Taleb, who has long been a critic of the unthinking use of Gaussian models in finance, I am sure realises this.

A plan for infrastructure projects

While on the subject of infrastructure, and the UK Government’s lack of any apparent action to start new infrastructure projects despite the economic crisis, here is a draft plan of action:

  • Start with a national competition for suggestions for new infrastructure projects.  People and businesses in regional communities have loads of ideas for projects – should anyone in Westminster bother to listen.   Perhaps allow 1 month for this, so Month 1 is spent soliciting proposals.  Creating a press release to announce the competition and a web-site to receive suggestions could be done within a day.
  • In the meantime (also during Month 1), create a temporary government agency like Australia’s national infrastructure agency, to receive these proposals and do a preliminary filtering in terms of (say): employment impact, wider business impact, social impact, cost, and long term potential for follow-on benefits.   A leading management consulting firm or two could be used to detail the criteria, assess all the proposals against the criteria (tedious but necessary work), and produce this long listing, winnowing down from (say) hundreds of proposals to (say) 50.  Month 2 could be devoted to this effort.
  • Then, have an appointed national committee, comprising politicians from all three major national parties, people from business and industry, the trades unions, people and politicians from the regions (say about 20 people) assess the 50 long-listed proposals and winnow them down to (say) 10.   This should be done in closed session in one, dedicated, all-day-and-all-night effort, over (say) 7 days.   We want the committee to bond, because we want their conclusions to be unanimous.
  • Then, prepare detailed technical and financial plans for each project on the shortlist.   This could be achieved within (say) 21 days.    As with the earlier stages, this work could be undertaken with the assistance of consulting and/or engineering firms, major corporations or banks  – there are currently lots of bankers at a loose end, I hear.  Hell, I’d even volunteer for this myself, because of the fun it would be and the importance of the work.
  • Then, fund the final 10 projects immediately and start digging ground (or spinning fibre, or whatever).  These projects should be give short, sharp names (eg, Fibre-up; Fast-Track) and short descriptors, so that every person over 16 can identify with them, and support them.  Insist that each team’s management produce detailed progress reports online each month, with (say) quarterly public hearings.   We want this work done, done well and done properly.

Total time, from start of campaign to shoveling: 3 months.
Of course, I realize getting major projects to shovel-ready normally takes longer than 3 months.  THIS FACT SHOULD NOT STOP ALL THESE PROJECTS STARTING SOMETHING WITHIN 3 MONTHS.  A key task will be creating semi-permanent, quasi-independent parastatal bodies (quangos) to run each project, to acquire land, employ people, etc.   That can all be done after the projects start, since the first main purpose of these projects is to boost aggregate demand and employment in the short run.    Our models here should be the  USA’s Tennessee Valley Authority and Australia’s Snowy Mountains Scheme, updated for the Internet age.
Not all infrastructure projects need to involve alteration to the earth’s physical landscape. My own proposal would be to create a major national organization – part-research lab, part-investment bank – to identify, to prototype, to seed, and to invest-in business ideas for future-generation Internet applications, starting from about Web 6.0 (whatever that will be) and upwards – a Xerox Parc for 21st-century e-services, with an investment budget of (say) USD 5 billion or so to start.  I would start this with public funding, with the aim of privatizing it once it becomes successful.
And (added 2009-02-12), if 3 months is too long (and it is), here are three potential major national infrastructure projects suggested by journalist Andrew Rawnsley:

  1. A national high-speed rail network (I would call this Fast-Track, or similar)
  2. A national, super-fast broadband fibre optic network (Fibre-Up), and
  3. A large-scale renewable energy production program, connected to the National Electricity Grid (Green-Power-to-go!).

There would be nothing stopping the Government spending (say) GBP 1 million on each of these to prepare outline feasibility and financial plans, with the aim of launching one of them within a month.
Building a national fibre broadband network without thinking also about what would run on it would not be sensible, which is why I propose the Web6.0 idea above.  But a little creativity could generate lots of proposals for non-physical infrastructure, which would create UK employment here and now, train people, stimulate demand, and leave something behind for future generations, for example:

  • Digitizing the contents of ALL Britain’s art galleries and museums, something which could employ artists, photographers, and lots of those unemployed media studies and IT graduates.
  • Digitizing the contents of the British Library, the main University Libraries and the national archives.
  • Digitizing ALL past census records.
  • Recording the life story of every citizen over 65.
  • Recording a performance at every live music venue in the country, including pubs and churches.
  • Producing online visitor guides to every locality in the country, annotated by people resident in the locality.
  • Producing a digital record (films, interviews, oral histories, photos, etc) of every factory facing downsizing or closure, with a record of the skills and networks being lost.

It should not need saying that all this digitized information, if paid for from the public purse, should be made freely accessible online.  These projects could be our generation’s equivalent of the Works Progress Administration.   I am sure there are many more ideas, both sensible and wacky, than these.
Well, Mr Brown?  What are you waiting for?  How about some vision?  If not these projects, then what?  If not now, then when?