Some men will tell you that it's your mind they're after. In the case of Christine Downton and the men from the military-industrial complex, it was true. Her mind contained expertise that researchers at Hughes Electronics - missile makers, robot designers, spy satellite pioneers - were intent on sucking out. Enemy secrets? Weapons plans? No, the nitty-gritty of financial markets.
In 1993, Christine Downton, a star analyst at London-based investment house Pareto, flew out to Hughes' labs in Malibu, California, to have her knowledge of the world's bond markets downloaded into a machine. Now that knowledge sits on a Macintosh at Pareto, looking after funds worth US$200 million. Another clone of Christine will shortly join it, looking after stocks and shares. Pareto and Hughes have between them decided that, in the war for the world's markets, the mechanised divisions are going to win.
Downton, Ron Liesching and the rest of the Pareto-Hughes team believe that their artificial intelligence (AI) trading system is one of the first concrete steps towards a shakeout of the financial industry that new technology will precipitate. Computer- based AI technology will automate many of the jobs of analysts and dealers, and destroy the closed shop of humans in the upper echelons of finance. City fat cats will see their value plummet; only those who embrace the technology will survive.
Plenty of Pareto's opposition will scoff at its RoboTrader, having followed AI's track record in the markets. Scientists have long seen the markets as providing problems tailor-made for their technologies - complex, with multiple variables and large volumes of data that must be processed rapidly. Financiers have dreamt of magic tools with which to make their fortune. As a result, lots of money and most of the contents of the AI tool box - expert systems, case-based reasoning, neural networks and genetic algorithms - have been thrown at the problem. But the results have been disappointing, and systems that start up in a blaze of publicity, like Citibank's neural network for foreign exchange trading, tend to end up with their plugs quietly pulled.
Liesching, Pareto's director of research, knew about the pitfalls before he embarked upon the company's robot trader project; he'd suffered through some of them while at County NatWest Investment Management. He knew from the start that such projects take time and money: more than a year and more than a million pounds. But he's not the sort of man to be deterred by that. He's as sweeping and startling in his predictions about technology's possibilities in financial markets as he is scathing about other people's failures to realise those possibilities.
In the early 1990s Liesching started hunting for techie partners who might help Pareto automate the stewardship of at least some of its £15 billion worth of funds. Bell Labs, Digital Equipment Corporation and Unisys were all found wanting. They had clever, powerful tools, but they didn't meet the requirements of the financial world - a peculiarly tough place to operate.
"There's a high data rate, there's a lot of noise in the data, there's errors, it's not all numbers, and you've got to do the job reliably; if you're wrong you're gone." Liesching's analysis makes it sound nasty, even hellish. Which is where the military comes in. War, after all, is also hell.
"The military deals with dirty applied problems just like you get in finance," says Liesching. He's not the first to spot the similarity. Sun Tzu's The Art of War does a brisk trade among business and finance types - as does the Marine Corps' Warfighting Manual. In fact, last year the Marines moved into the New York Stock Exchange, putting officers under training into the trading pits. You can see the similarities to the modern command post: lots of information but not necessarily enough, lots of decisions and a lot riding on it all. According to General Richard Hearney, assistant commandant of the Marine Corps, they wanted to compare how the two professions dealt with the types of stress normally associated with the battlefield.
The similarities explain why both soldiers and financiers are eager to use AI. They worry about information overload; they also worry about emotional stress. Emotions, in Downton's view, are the rational trader's enemy. "Emotions distort people's rational judgements. There's a fear factor in managing money - people tend to make mistakes when they're losing money. They also make mistakes when they've made money because they get big-headed."
There are other human irrationalities, too, those that Downton calls "cognitive biases". "Very often, in the short term at least, the market will get fixated on one variable rather than a whole range. Or people get hung up on the most recent piece of information they received, or a distorted assessment of the information - because human beings just have processing limits."
Processing limits that are becoming ever more of a hindrance. Consider the research finding that people can process only around seven chunks of information at any one moment. Twenty years ago, when an analyst typically looked at only a few pieces of information in three or four bond markets, this didn't matter; now it does. "If you want to compete in this sort of product you are probably going to have to cover about 10 to 15 markets," says Downton. "You might want to look at, say, 10 to 20 variables for each of three sources of return for 10 to 15 markets. You're looking at trillions of potential combinations."
Anyone who's tried to make sense of a Tom Clancy novel will know that the modern military is similarly complicated, which is one of the reasons that armies spend quite a lot on AI. Many of the key university AI labs were started - and are still funded by - the US government's Defence Advanced Research Projects Agency (DARPA), and their research has spawned many technologies that have since been used on the battle- field. The image processing techniques of machine vision, for example, have been used for analysing images from satellite cameras or radar, sonar or infrared sensors. Guided-missile developers have adapted tracking and path-finding algorithms originally written for laboratory robots. Even the age-old chore of calculating troop movement logistics has benefited from problem-solving and expert-systems programs.
The Gulf War most vividly displayed AI's usefulness in battle in 1991. The "smart" bombs were not that smart - they mostly just homed in on splashes of laser light. But DART (Dynamic Analysis and Replanning Tool), a distributed-planning program developed by AI veterans at BBN Systems and Technologies, was very smart indeed, and proved an invaluable help in sorting out the scheduling nightmares of an operation as vast and sprawling as Desert Storm.
It was this background that Hughes brought to the table. It also brought an eagerness to diversify from what appeared to be a shrinking defence market. The fit with Pareto seemed perfect, and quickly developed into a real partnership. All that remained was to show that AI really could master the trader's art.
Uploading DowntonAs a teacher of that art, Downton would be hard to better. She has studied the markets for 20 years as an academic and practitioner, including spells with the Bank of England, the US Federal Reserve Bank and County NatWest Investment Management. That's a lot of experience. And it's coupled with a certain individual flair. Liesching remembers vividly their first meeting - among a bunch of City gents in suits, Downton cut a striking figure with her bright red hair, jeans and motorbike.
The man at Hughes' research labs in Malibu assigned to squeeze out Downton's experience was Charles Dolan, who has a PhD in cognitive psychology and computer science from the University of California, Los Angeles. Dolan likes to devote himself to "world-class-hard problems". To begin with he wasn't sure finance offered any; Downton convinced him otherwise. And the project held greater appeal to him, too. As Dolan points out, "In the military it takes 14 years to develop a new missile before it goes into production. By that time you don't see much of your technology because it goes through so many transformations. In finance you see it right away."
Dolan's approach is a mixture of traditional symbolic logic approaches to AI and newer connectionist theories where intelligent behaviour emerges out of the development of an artificial "neural net" rather than being programmed in from the start. Dolan's view is that the two are part and parcel of each other - that within the brain's networks of neurons there is structure, and that this structure is the embodiment of symbols. Dolan tries to create such knowledge spaces on the computer based on the symbolic structures that have been laboriously built into the wetware of his willing subjects.
One of the limitations of the shells in conventional expert systems - the software that provides a framework for encoding expert knowledge - is that from the start they impose a rigid way of reasoning. Dolan's M-KAT (Modular Knowledge Acquisition Toolkit) recognises that different experts have different symbol spaces, and that their reasoning follows from the details of the space in which it takes place.
Because knowledge engineering means cross-examining the expert's thought processes, it often exposes charlatans. Downton proved to be the genuine article; indeed, "she had quite a bit more access to her internal thought processes than most experts do," says Dolan. Still, it took a gruelling series of sessions spread over 18 months to get a fair sample of those processes into the box, with Dolan switching from tool to tool to find ways of mimicking the trains of thought that Downton described.
The most difficult part of the engineering process was capturing Downton's "feature extraction" abilities. "When I look at a variable, I ask a number of questions about it, such as, is this inflation number high? has it been high for long? and what are the recent trends? The most time consuming part was explaining what I meant by 'high', and then helping them design something that would look at a particular number and come up with exactly the same assessment as I would."
The hard work paid off in real benefits for Hughes. "I expected most of the benefits to be much more 'squishy' ", says Dolan, "where a researcher comes away with a new perspective on a problem." In fact the process left the team with knowledge acquisition techniques a whole order of magnitude faster than their originals. Those are now being used in the US to build simulations of entire theatres of war. "They want to simulate all the entities with great fidelity, down to the individual combatant. Unless you have very good knowledge-acquisition techniques on tactical decision making, you can never have hoped to gather enough knowledge to get you that fidelity of simulation," says Dolan.
For Pareto, the result was a system of 2,000 rules called the Global Bond Allocation Strategy. The system takes in around 500 items of economic information a month - things like countries' public-sector and current-account deficits, inflation rates, money-supply figures and so on - from electronic market-data feeds. After crunching through millions of permutations of the information and potential deals, it outputs its conclusions as a series of recommendations of bond trades, such as selling holdings in Denmark and buying bonds in Germany. The recommendations are passed to a flesh-and-blood trader at Pareto who then makes the deals.
A silicon chip off the old blockVilfredo Pareto was a 19th century economist who pioneered the introduction of higher mathematics to economics. The company that sports his name is, fittingly enough, devoted to a "quantitative" approach to trading - jargon meaning that all its trading and investment is done using models, albeit simplified ones, of what is actually happening, rather than feelings and theories about why it is happening. As such, it seemed natural for Pareto to turn to AI - and AI fit into the firm easily.
So how has the RoboTrader performed? In the markets, the rate of return from trading is a function of risk: the more profit you want, the bigger the risk you must take. Pareto manages money for major public and corporate pension funds. Pension funds are generally conservative - which means they want low risks and must settle for lowish returns. At the moment the system is mostly managing portfolios with a low risk level - around 2%. On these, claims Liesching, the system produces returns between 1.5% and 2% above a benchmark bond index. This just-better-than-one-to-one risk-to-return ratio is the kind of solid-but-not-startling performance that large pension funds seek.
The returns are not startling. But then, the RoboTrader isn't being asked to startle; the low risk levels are part of its (reprogrammable) parameters. And they are all the program's own work. Downton resists any temptation to override the system's recommendations, especially when the markets are volatile - that would defeat its whole purpose. "Few people are prepared to rely completely on analytical processes - they want to second-guess them in some way. That's when their emotions get involved. And it's probably just when they should be relying on their models that they are throwing them out of the window." Though Downton and her silicon twin are nearly always in agreement, " ... sometimes there are slight nuances between what it recommends and what I think I would do. But when I look into it I see the machine is right in that it has noticed information I hadn't remembered, or it's more detached."
One company that thinks it has seen the future at Pareto is Bermuda-based insurance leader Exel. It liked the robot so much that it bought the company, recently taking a 30% stake in the firm Pareto Partners with the intention of merging some of Pareto's investment and risk-management methods into its insurance products. According to Exel Vice President Gavin Arton, the company plans to try the Hughes-Pareto knowledge-engineering techniques to automate some of its underwriting expertise.
And Pareto is furthering its own commitment to AI "wherever appropriate". Shortly after the bond machine got up and running, Downton went back to Hughes for another bout of brain-draining, this time to extract her expertise in equities and their inter-relationship with the bond markets. From this the Hughes-Pareto partnership has built a second knowledge-based system - its Global Asset Allocation Strategy. The system is now undergoing final testing, with the firm trading its recommendations on paper to see how they would do. Pareto is confident it will be ready to go live with real money on September 1st and has already signed up a customer with a $50-million portfolio.
Others remain to be convinced of the success of the existing model, let alone the new one. One Pareto client directs a pension fund for one of the US's biggest technology companies (like most of Pareto's customers, he asks not to be identified); he points out that investment is not quite the same as conventional scientific problems. "You are part of the problem you're trying to solve. If you build a system that's really good and picks profitable bonds, then the very fact that you buy those securities affects the markets. When you're managing $15 billion, your actions can move the markets so they don't work as they did in your solution. There's a feedback loop that causes your solution to become part of the problem."
Liesching is not too worried. He believes that AI and agent technology will cut a swathe through the industry, automating thousands of jobs or downgrading their skills, not necessarily because their results are that much better, but simply because they're cheaper. "People in finance are generally over-paid and under-qualified, and there's too many of them," he says.
Most of what these people - analysts, strategists, marketing executives and so on - do is what Liesching calls "knowledge-directed searching". "People used to do it in the past, but now because of the vastly increased data flows, that's becoming impossible." Downton says that no human expert could process the volume of information that the Global Bond Allocation machine sucks up with the same efficiency and consistency. In the end, Liesching believes, such systems will lead to radical downsizing in the City, and will burst its inflated-salary bubble. One by one, the functions people currently perform, and for which they charge huge margins, will be picked off and automated: identifying arbitraging opportunities, building and optimising portfolios, brokering, trading and managing market risk. The Internet will hasten the process as it provides a low-cost channel via which to deliver sophisticated services directly to the consumer.
Liesching's predictions seem to fly in the face of current trends, where human financial expertise has never been at a higher premium and the direction of City salaries appears to climb relentlessly. But he is adamant that a major shakeout is coming. "Whoever can replace these people with machines will win - even if the machines are only half as good - because they can work 24 hours a day and don't have the personality side-effects of some of these highly paid individuals."
Downton has no worries that her clone will take over her job. She finds it "... enormously liberating. It releases the human expert from the drudge work of information processing." It lets her spend more time thinking about the markets and less time immersed in them. "We believe the best way you can use human insight is in designing models, not in second guessing them."
It also gives her the time to look for changes in the way they operate - changes that the AI would not be able to spot. As John Maynard Keynes remarked, when the facts change, it's time to change your mind - and Downton now has two minds to change, with a third on the way. As yet, though, the only changes in the market have been superficial ones, with which the system's learning algorithms are perfectly capable of coping.
The machine may mimic an expert, but it isn't one; Christine Downton, capable of changing her minds, real and prosthetic, is. That still gives her, and true experts like her, the edge. In the long run, the technology might capture the gift of developing expertise, or even cut away the need for it. After all, if all the traders are rational robots - not emotional, cognitively biased people with worries and fears and vanities - the markets might behave more efficiently, removing many of the cunning possibilities for arbitrage that experts can discover. Until that day, there's money to be made.
Clive Davidson writes about science and technology, mostly for the financial press.