Risk Management

11:08 AM
Melanie Rodier
Melanie Rodier
Commentary
Connect Directly
Facebook
Google+
LinkedIn
Twitter
RSS
E-Mail
50%
50%

Watson, From Jeopardy To Wall Street Hero? Hmm. Maybe.

Could Watson have helped prevent the financial crisis?

The debate of Man Vs Machine - whether one day computers will exceed human intelligence - has been raging for decades. It has also increasingly permeated Wall Street board rooms with questions over whether one day algos will completely replace traders.

Earlier this week, at the High Performance Computing on Wall Street show, I met Eddie Epstein, manager of unstructured information at IBM Research - one of the chief engineers behind Watson, the super computer who beat the champions on Jeopardy! a couple of months ago.

Watson was built to compete at the human champion level in real time on Jeopardy. Its computing power can be compared to over 2,880 computers with a single processor core linked together in a super high-speed network. The hardware that powers Watson is one hundred times more powerful than Deep Blue, the IBM supercomputer that defeated Gary Kasparov, the world's greatest chess player, in 1997.

Watson made Jeopardy history on a cold February night. Well, one might think, a computer should win a quiz over a human being. After all, when you want to know the answer to just about any question nowadays, you wouldn't think of actually asking someone. You just Google it, right?

Well, according to Epstein, what's truly amazing about Watson is that it is the first computer able to analyze both structured - or factual data - and unstructured data. It can parse natural language, understand opaque statements, and come up with intelligent answers.

But perhaps the most interesting aspect of Watson on Jeopardy! was the fact that it showed how confident it was about the answers it came up with. If the super computer was really confident, it would buzz in its answer. (average response time = 3 seconds). If Watson hesitated, maybe the human contestant would buzz in their answer first. In some instances, Watson decided it shouldn't risk choosing an answer at all.

Hmm. So given that financial models seemingly contributed to a false sense of security on Wall Street that encouraged firms to scale up risks, if Watson is humble enough to know when he doesn't know the answer, could he have helped prevent the financial crisis? Imagine if an alarm had gone on in Wall Street back offices in '07 and computers said, "Hang on guys, I'm going to skip this one. I just don't know what the right answer is."

Epstein, who was responsible for literally getting Watson up to speed, and Jean Staten Healy, director, cross-IBM Linux, IBM, argue that Watson could have a huge impact on the financial industry.

Hedge funds for example could use the computer's powers for research. Watson could help with arbitrage trading or to define acquisition targets, they say.

According to its makers, Watson could help prevent another financial crisis since it could analyze in real-time all the interconnecting factors that eventually brought the economy crashing.

Watson could indeed be useful for research. It can analyze petabytes of information, and therefore could come up with an answer as to which companies, for instance, are worthy of acquisition. It could also show how 'confident' it is about its decisions by demonstrating how it came up with its answers.

But this raises some questions. Firstly, someone would presumably have to spend the time entering into the system these petabytes of information in the first place. Who would do this? How long would it take? And how do you know exactly what information to put into the system?

Secondly, what if everyone else on Wall Street is also using Watson? (or at least the big firms who can afford the millions of dollars Watson would presumably cost). Your competitive advantage would be annihilated - or at the very least, Watson would perpetuate the herd mentality. (I'll do whatever he's doing, thank you very much).

There are other shortfalls. Adam Honore', director of research at Aite Group, points out that a computer can't account for subtle nuances that often come into play when you're talking about human interaction.

Watson may come up with a list of worthy M&A targets, but what if there is bad blood between executives at the company that a firm is looking to acquire? Surely that could taint a decision to buy another organization and its successful outcome.

There can also be such a thing as information overload. Honore' points to the theory exposed by Malcolm Gladwell in his seminal 2005 book, "Blink: The Power of Thinking Without Thinking."

Sometimes having too much information can interfere with the accuracy of a judgment, or even a doctor's diagnosis, according to Gladwell's theory. Having too much information can be irrelevant and cloud the decision maker's judgment.

Still, if Watson is used properly as a research tool rather than an idea generator, the benefits to Wall Street could be huge.

That is, as long as someone physically looks at the results that Watson comes up with, and ultimately, a human makes the final decision.

Melanie Rodier has worked as a print and broadcast journalist for over 10 years, covering business and finance, general news, and film trade news. Prior to joining Wall Street & Technology in April 2007, Melanie lived in Paris, where she worked for the International Herald ... View Full Bio
Comment  | 
Print  | 
More Insights
More Commentary
Micro Data Challenges in an Era of Macroprudential Regulation
Research and statistical analysis experts at central banks are tasked with developing sophisticated forecasts and models to identify systemic risk. Yet they are spending most of their time acting as data entry clerks, rather than developing these models.
The Perks of 'SmartSourcing' Shared Services in Financial Industry
A breadth of vital but undifferentiated business processes are still being replicated across the industry. They are all candidates for centralization.
Managing Social Media Risk Strategy: Technology Can Only Go So Far
Advanced analytical technologies are an important part of a social media risk management strategy, an Accenture report says, but the technology must be balanced with training and procedures.
Cross-Asset Universal Product Identifier: The Solution the Industry Is Looking For?
A UPI will enable a holistic approach to identifying all trades and positions. While such an idea sounds great in theory, historical attempts at achieving global agreement have fallen short.
Register for Wall Street & Technology Newsletters
White Papers
Current Issue
Wall Street & Technology - Elite 8, October 2014
The in-depth profiles of this year's Elite 8 honorees focus on leadership, talent recruitment, big data, analytics, mobile, and more.
Video
Inside Abel Noser's Trading Floor
Inside Abel Noser's Trading Floor
Advanced Trading takes you on an exclusive tour of Abel Noser's New York trading floor, where the agency broker known for transaction cost analysis, is customizing algorithms for the buy side, while growing its fixed income trading and transitions business.