Machine Learning: Hedge Funds and Quantitative Trading

The article was written by Jordan Klotnick, a Financial Analyst at I Know First. He graduated from Monash University with a Bachelor’s in Business – Majoring in Marketing.

Machine Learning

Summary

  • Artificial Intelligence and Hedge Funds
  • Hedge Funds Train Computers To Think Like Humans
  • Deep Learning Difficulties
  • Artificial Intelligence
  • I Know First and Hedge Funds

Artificial Intelligence and Hedge Funds

AI hedge fundA hedge fund’s main purpose is to maximize profits for its investors, while managing the risks. A hedge fund is a partnership between fund managers and can be seen as more of a pooling of funds by only certain investors. These investors have to meet certain qualifications, one of which is having a net worth of over $1 million and are referred to as “sophisticated” investors. The idea of pooling funds is that investors can invest in a specific company, accompanied by other companies, that will essentially safeguard the investment as a whole. When the market goes down, hedge funds aren’t affected as much because of this “insulation.”

Just like in many other areas of finance, hedge funds are becoming computerized. Famous hedge fund managers are being replaced by super computers with complex algorithms to watch and predict the market in order to make better bids. Rather than these managers working with hedge funds, PhD holders in mathematics are developing new ways to look at data through mathematical equations, essentially replacing the hedge fund manager.

More recently t47629446-a-business-person-is-writing-down-math-formulas-on-the-glass-screen-in-the-evening-modern-panoramiche hedge fund industry “has suffered the biggest quarterly outflow since the financial crisis,” because more investors trust supercomputers than seasoned and skilled hedge fund managers (retrieved from The New York Times). Statistics show that this year, Investors have invested about $7.9 billion into quantitative hedge funds. Companies’ AUM now using the technology have more than doubled from $408 billion to $900 billion in seven years, according to Hedge Fund Research.

Hedge Funds Train Computers To Think Like Humans

Artificial intelligence technology called deep learning is mimicking the neurons in our brains, holding out promise for firms. WorldQuant is supposedly using it for small-scale trades. There are a few companies that are going to take this on too, with Winton and Two Sigma also coming into the game.

The firms are hoping that this artificial intelligence will help them give them an edge in the technological arms race in global finance. If this is correct, neural networks could help propel the transformation of finance, putting human against machines, therefore threatening old school jobs.

“Having witnessed in the 1990’s the hype and subsequent failure of hedge funds purporting to use neural networks, we tend to be skeptical of claims that ‘deep learning’ will solve the general problem of investment management,” said Winton, a multi billion dollar quant firm in London.

Quant funds are following tech giants leads, which have proved to be the mettle of deep learning. The technology, involves super-powerful computers and data to do its job. It’s believed to be about five years away from becoming a mainstream tool at hedge funds, according to Nicolas Chapados, a computer scientist.

“There’s a huge class of deep-learning models used in tech firms that can be adapted to financial processing,” said Chapados, who co-heads quant fund Chapados Couture Capital in Montreal and Element AI, a research firm that uses the technology.

Deep Learning Difficulties

Chief scientist at Man AHL in London, Anthony Ledford, said his researchers have spent over a year developing deep learning. The firm hopes for this to start live trading soon. The hedge fund has been allocating money to a machine-learning strategy.

“In deep learning, one doesn’t have to define the features in advance, deep learning aims to identify them for you,” said Ledford, whose unit of Man Group Plc manages about $18.3 billion. “We know a lot of potentially predictive features already given AHL’s three-decade history, so the real challenge is often shaping the methodology, data and computations so we discover new features, not the ones we know already, which is very challenging in itself.”

Artificial Intelligence

It used to be that if one wanted a computer to carry out an action they would have to program it to do that specific action. This took an excruciating amount of time as we would have to tell the computer to do exactly what we wanted it to do and it was unable to carry out an action, which we were not able to do or to tell it to do. Programs had no independent intelligence and could not make decisions by themselves.

In 1956, computer gaming pioneer, Arthur Samuel, wanted his computer to be able to beat him at chequers. Samuel then programmed his computer to play against itself thousands of times to the extent that the program accumulated sufficient knowledge of the game. By the 1970s, his program was proficient enough to challenge and beat the masters. Arthur Samuel is therefore credited with being the pioneer of artificial intelligence (AI).

I Know First and Hedge Funds

I Know First Algorithm predicts a growing universe of over 7,000 securities for the short, medium and long term horizons daily by applying Artificial Intelligence and Machine Learning techniques to search for patterns and relationships in large sets of historical stock market data. Through its self-learning ability and flexible multi-layered neural networks structure, the algorithm is able to learn from, adapt to and evolve together with continuously changing markets. It offers an independent, objective and unique perspective on the financial markets and doesn’t rely on any human derived assumptions or traditional theories and models that often do not hold (any more).

The results of intense learning and prediction cycles are aggregated into two indicators per time frame: signal and predictability. While predictability indicator helps to identify and focus on the most predictable assets, the signal is used to define and rank the trades and is related to the magnitude of expected return.

The applications of the algorithmic AI-based forecasts are multifold, and use empirical prices of the stock market to create daily predictions.

The scalability of the algorithmic predictive system allows I Know First to offer custom forecasting solutions to hedge funds and other financial institutions, so they can identify the best opportunities as discovered by the self-learning algorithm within the investment universe of their interest. Further, the solution can be used as a decision support system in form of an algorithmic screen integrated into client’s investment process in order to confirm or reject investment ideas before the execution.

Moreover, I Know First develops and back-tests systematic trading strategies, which are used in partnerships with, hedge funds and other asset managing entities. These strategies are rules-based and utilize algorithmic forecasting indicators mentioned above in order to rank and select the trades as well as time the execution. The type of strategies varies, including mean-reversion logic and more trend focused approaches, all generating high positive alpha while keeping beta in the 0.3-0.8 range, yielding overall high risk-adjusted returns. The strategies can be used in partnership with I Know First to launch hedge funds, mutual funds or other investment vehicles.

Conclusion 

The future is very unexpected, especially when it comes to jobs for investment managers. Looking forward, there are many more advancements in technology, specifically in the artificial intelligence sector. Thus, it will take away from individuals jobs. There is a lot of uncertainty how accurate this may be, but only time will tell.

 

Quantitive Trading