Discovery and Exploration

How fund managers can apply AI to turn data into insights, reduce bias in decisions and generate alpha

Share this post:

Key Points:
– In this age of rampant data growth, the only way to reliably beat the market on a risk-adjusted basis is to mine unstructured data faster and more accurately than your competitors
– Investment firms that combine machine speed, scale and accuracy with human creativity, nuance and contextual awareness will capture the lion’s share of alpha
– The investment management business is at the dawn of an era of super-intelligence defined by a partnership between man and machine
– Those firms that adapt will not only find alpha with greater speed and accuracy, they’ll literally change the game of managing and growing investments

Create your Insight Engine today

 

Overcoming cognitive bias to generate “alpha” in the digital age

Traditionally, funds have relied on asymmetrical informational edges to outperform the market – interpreting news, research and filings faster than others, or rapidly leveraging deep industry knowledge and expert networks, etc. But as unstructured data – such as texts, tweets, images, videos and sensor data – continues to grow faster than we can possibly keep up with, it’s become almost impossible for human beings to consistently make sense of financial markets.

This deluge of data causes inefficient information processing and reasoning, which leads to biased investment decisions and increases risk exposure.

Many market participants still rely on antiquated tools and methods like price charts and asset diversification to find opportunities and manage risk in financial markets. Traditional methods on their own can’t keep up with the ever-increasing flood of data, let alone the force-multiplier of crowd psychology.

The digital revolution has altered the very nature of market volatility, propelling the cognitive biases of investors into manias for no apparent rhyme or reason. Through smartphones and social media, we’re imprinting almost every impulse into the cloud. It’s no wonder that asset managers increasingly struggle to parse signals from noise and beat benchmark indices.

In this age of rampant data growth, the only way to consistently generate alpha – or reliably beat the market on a risk-adjusted basis – is to mine unstructured data faster and more accurately than your competitors do.

It’s just a matter of finding deposits of alpha within the tsunami of data we encounter every day. You need a deep-data drilling rig – a platform built upon the symbiosis of human and machine rooted in the idea of amplified intelligence.

Data, data, everywhere—but where are the actionable insights?

According to The Economist, the world’s most valuable resource  is no longer oil, but data. It’s hard to wrap our minds around it, but IBM Research estimates that over 90% of the world’s information was generated in the last few years. Most of that information is dark, or unstructured, in the form of research, news articles, social media posts, and comments that traditional computing systems can’t interpret.

IDC estimates that by 2020, we’ll be generating data at the rate of approximately 1.7 MBs of unstructured data per second for each person on the planet. That translates to about 14 million GBs of unstructured data each second in less than 3 years. Clearly, the explosion in unstructured data is outpacing our human processing capabilities.

It’s hard to imagine a future in which any knowledge worker –  especially those working to make sense of financial markets – can make good decisions without the aid of contextually aware machines that learn.

AI that understands the nuances of financial markets — at scale

Although machines can continuously read massive amounts of information without taking a break or getting tired, they can’t yet natively understand the meaning or context of that information without our help. On the other hand, even if humans are able to teach the machines to understand the basic context of a particular domain, contexts evolve and change over time and humans just aren’t equipped to keep up.

For example, imagine that a human trader who has acquired knowledge and experience sees an article flash across her Bloomberg terminal with the headline “Tesla is going to the moon!” She may interpret the phrase to mean that Tesla’s stock price is rapidly rising – a positive statement.

But how would a machine interpret the headline “Tesla is on fire!”? Would a machine understand that “Tesla is going to the moon” and “Tesla is on fire!” mean the same thing – a rapidly rising stock price? Or, in another context, would it be able to reason that “Tesla is on fire!” means flames have engulfed an electric car – a statement that could actually have a negative impact on Tesla’s stock price?

The challenge is to interpret nuance at scale, as context and markets evolve over time.

Machines that parse sentiment in vast volumes of information without considering the challenge of context – or nuance – only exacerbate biased decisions rather than correct them. Generating alpha in the digital age is a matter of teaching machines how to speak the language of a particular financial market domain and scaling contextual awareness through the use of deep learning.

So, what will the new tools – the deep-data drilling rigs – of the investment management future look like?

They’ll be intelligent, which means they’ll be able to reason, understand, learn and interact at scale, augmenting human reasoning and learning beyond our biological abilities.

They’ll be contextually aware and adaptable, able to interpret new information according to specific market domains such as global macro, long/short equity, value, merger arbitrage, fixed income, event driven, etc.

And they’ll work with human beings, helping us cut through data clutter to uncover answers to complex questions and make better decisions confidently and with speed.

Those investment firms that combine machine speed, scale and accuracy with human creativity, nuance and contextual awareness will capture the lion’s share of alpha. Those firms unwilling to adapt to the new paradigm will fall prey to the market inefficiencies that others will take advantage of.

Today’s explosive growth of unstructured data marks a critical turning point in human evolution. It’s going to become increasingly challenging to make sense of the world without the aid of machines. You could say that the very nature of human intelligence is being redefined, and thriving in this digital revolution will be a function of allocating cognitive resources to tasks that optimize our innate curiosity and creativity. In fact, at the root of all alpha are good questions.

The promise of artificial intelligence is that machines will free up cognitive bandwidth currently spent on laborious tasks such as listening to endless earnings calls to track changes in key topics over time and across sectors versus market chatter and expectations. Or continuously scouring the web looking for clues to verify acquisition rumors, and looking for clues across thousands of disparate data sources as to how the Federal Open Market Committee (FOMC) may adjust interest rates in the near future.

“Leveraging A.I. technology,” says Marc Lebensfeld, an experienced buy-side research analyst, “remains an untapped source that can unlock valuable hidden data for portfolio managers, reducing research time and speeding up the process to make investment decisions.”

Will your fund be disrupted, or will you be the disruptor?

Frankly, asset managers who continue to employ traditional methodologies and tools to make sense of all this information are facing nothing less than an existential crisis. They have three options.

  • Ignore the explosion in unstructured data, continue to make increasingly biased decisions and ultimately take too much risk
  • Return capital to investors and close up shop
  • Adapt, which may not be as difficult as it sounds

Consider how the oil industry has evolved. Oil used to lie just below the surface, easily extractable using rudimentary tools. Today, extracting oil is a complex process that requires sophisticated tools like deep water drilling rigs outfitted with de-gassers, de-sanders, reciprocating piston/plunger devices, and much more.

Like oil, alpha used to be easy to extract by analyzing information that fit neatly into rows and columns of spreadsheets and structured databases. Today, alpha is buried deep under volumes of messy, unstructured information.

Today’s investment managers need new tools that can address this changed environment. Tools that can rapidly sift through and make sense of financial news, earnings calls, analyst reports, government filings, video interviews, industry blogs and market chatter, at scale, to distinguish bias from fact, trend from hyperbole and enable asset managers to discern real risk and value.

“Markets feed insatiably on information,” explains Matt Addison, former Managing Director of Platinum Capital Management (UK) Ltd. and merger arbitrage practitioner. “They’re starving for value-sensitive unstructured data.”

The investment management business is at the dawn of an era of super-intelligence defined by a partnership between man and machine. Those firms that adapt will not only find alpha with greater speed and accuracy, they’ll literally change the game of managing and growing investments.

That’s why at Accrete.ai, we’re working to forward this promise for investors. We orchestrate networks of independently sourced domain experts who don’t know each other and who collaborate to seed natural language classifiers with bias-free contextual awareness. And we employ deep learning algorithms that empower machines to reason, understand and learn at scale, leveraging IBM Watson for high throughput computational infrastructure.

 

Learn how your business can create insight engines and disrupt your industry with Watson Discovery Service.

Add Comment
No Comments

Leave a Reply

Your email address will not be published.Required fields are marked *

More Discovery and Exploration Stories
December 12, 2017

3 types of business chatbots you can build

No matter what type of bot you decide to build, it is important to give your bot some life and personality, make it useful, and make sure it’s easy to use. Our blog highlights the top three chatbots you can build for your business.

Continue reading

December 1, 2017

The future of AI Revenue: Top 10 use cases for AI in the next decade

6 AI segments will account for a significant change in revenue for AI enterprise software. These functional areas are applicable to many use cases, industries, and generate benefits for both businesses and individuals. Here are the top ten use cases which will reap financial rewards for AI technology product and service companies, and a broad spectrum of benefits for everyone else.

Continue reading

November 16, 2017

Get the Forrester Report on how to make customer service smarter, faster and more cost effective

Call centers executives face the constant challenge of meeting customer expectations and business cost goals. Customers prefer interacting with virtual agents, and are choosing messaging over phone calls for issue resolution. This Forrester report outlines the trends that will enable call centers to become smarter and more strategic.

Continue reading