Machine Learning Hedge Fund: Artificial Intelligence, Algotrading and Hedge Funds

taliTali Soroker is a Financial Analyst at I Know First.


Machine Learning Hedge Fund


  • What are hedge funds
  • Hedge fund strategies
  • Quantitative Analysis and Artificial Intelligence
  • Bringing AI based trading to the masses
  • I Know First Algorithmic Hedge Fund Solutions

Hedge Funds

At the end of 2006, a sizeable hedge fund, Amaranth Advisors was taken down by a wrongly placed bet on the weather. The fund lost $6 billion dollars after a mild winter led to rapidly decreased natural oil prices. Hedge funds operate based on high risk, high return trades but this also means little transparency for investors and the risk of one bad call leading to the loss of billions of dollars. Hedge funds are private investment funds that work with higher risk trades that bring much higher returns than mutual funds or other investment entities. These funds are only open to sophisticated investors, either “accredited investors” or “qualified clients” depending on the type of the fund.

The first hedge fund was formed in 1949 by Alfred Winslow Jones. He employed two strategies with his fund, short-selling and leveraging. Using a combination of these strategies, Jones secured sizeable returns in the first few years. These tactics are still used by hedge fund managers to secure millions of dollars in returns.

Today, there are upwards of 10,000 hedge funds that are managing an estimated $3 trillion in assets. Despite the high risks involved and the impact that their investments have on the market, the SEC has not strictly regulated the operations of hedge funds. This is largely what is contributing to a lack of transparency for investors who can’t see how their money is being managed by the head of the fund that they’re investing with.

Machine Learning Hedge Fund

Hedge funds started out employing just two strategies, short-selling and leveraging. As the market changed, these funds developed to become more flexible in their investment options. Today, funds are using a wide variety of strategies such as equity hedging, achieving market neutrality, relative value arbitrage, and convertible arbitrage to name a few.

Hedge fund managers look at market data in two different ways. There are funds that are run using fundamental analysis and there are funds that are run based on quantitative analysis. The fundamental analysis aims to use market research on the value of different securities and determine which assets are “undervalued” and “overvalued”. Conversely, quantitative analysis involves the use of complex mathematical formulas and computer models to create time-series models of the market in order to identify ideal short and long-term positions on different assets.

Quantitative Analysis and Artificial Intelligence

Quantitative analysis in its static form is the use of computers to model the current market with models created by traders and mathematicians. These models, however, are not useful for extended periods of time because the market is constantly changing and shifting. In order to get accurate results with this type of quantitative analysis, the model must be formulated by hand and regularly updated to reflect changing market conditions.

In recent years, the interest in artificial intelligence has increased immensely likely due, in part, to the advancements that have been made in the field, specifically deep learning which involves training a large virtual neural network to recognize patterns in data and analyze the results. While these AI systems must initially be crafted by humans, they are then able to adapt to changing circumstances on their own. These systems are incredibly attractive to trading firms as they are always looking for new strategies to increase returns and decrease risk.

The amount of information that these systems can digest is almost unimaginable. They take in information from news stories and posts from social media sites like Twitter and Instagram to identify patterns and connections in the data. They can then make predictions about the direction of the market and in the case of many firms, they can then make the trade without the involvement of any humans.

More and more large trading firms are heading in the direction of AI. Hedge funds that are known for relying on data analysis like Two Sigma and Renaissance Technologies have said that they rely on AI now, a San Francisco start-up Sentient Technologies has been trading using an automated system for the last year, and firms like Aidyia are continuing to work on developing intelligent software that can function wholly independently from human interaction.

Aidyia is a large firm based in Hong Kong that just began trading on U.S. equities using an automated intelligence system. The company’s co-founder and chief scientist, Ben Goertzel, spent years researching and working with AI and cognitive science before applying them to financial markets. Aidyia now has a working AI system that can take in an inordinate amount of data that is incomprehensible to the human mind and return a simple prediction on the market. Executives at this company have no doubt that these types of automated intelligence systems will be running asset management in the near future.

According to Preqin, a well-known provider of financial industry data, there are about 1,360 hedge funds that are currently making a majority of their trades with help from computer models. These funds are managing about $197 billion in total. Additionally, approximately 40% of funds that opened last year use “systematic” trading.


It’s difficult to pinpoint exactly how successful this form of trading is in comparison to other forms, as there is little regulation in this new trading form. Aidyia has conducted extensive back-testing on their algorithm and CEO Ken Cooper claims that it has seen an average 25% year-on-year return. While good results in back-tests do not necessarily mean that the algorithm will be successful in the future, on their first trading day Aidyia reported a 2% return on an undisclosed amount. It’s not the most impressive return a company has had in one day, but it could signal the start of a new trading era.

While hedge funds dominate the artificial intelligence sector within the finance industry, companies like I Know First are breaking into the industry to provide this kind of technology to investors that can’t access it otherwise. This includes hedge funds and individual investors that are not a part of a large trading firms, as well. I Know First has a uniquely customizable solution for hedge funds that enables the fund manager to choose their preferred parameters. The algorithm that I Know First co-founder Dr. Lipa Roitman has developed is making daily forecasts on more than 3,000 markets over six different time horizons, from 3 days to 1 year. In 2015, an I Know First portfolio outperformed the S&P 500 yearly return by 96.4%. It’s unclear how many hedge funds are using AI now to boost performance and how many firms are successful. There is no doubt, though, that AI is changing the way the stock market works and the way traders will make money in the future.