Jacopo Credi, VP of AI at Axyon AI
Analysing market trends is fundamental to making effective investment decisions, so it is no surprise that technology has become integral in helping businesses make the right choices. However, advanced data analysis is constantly evolving and there is still some way to go before it becomes knitted into the fabric of every business.
The next challenge for investment analytics is predicting so-called ‘black swan’ events, which tend to catch businesses, markets, and governments by surprise. Fortunately, this is an area in which Artificial Intelligence (AI) is already making major inroads.
Sophisticated AI algorithms currently use data from past events as building blocks to predict industry trends and market shifts. As a result, this technology enables businesses to take investment analysis to a whole new level, since the more historical information the machine has access to, the better the analysis produced.
While this innovation does not eliminate the impact of a black swan event on the investment landscape, increased understanding of the marketplace can be vital for gaining fresh insights and predicting how different sectors will respond to different situations. When used in this way, AI can therefore help to limit the shock of a major incident dramatically.
A new hope
Despite the perception that black swan events are seemingly impossible to anticipate, incidents ranging from political unrest to revolutions in technology typically have a number of trends in common. As a result, the outcome of previous events can be used to piece together a more accurate image of what the market’s reaction would be if something equally unique, but with similarities to previous events, were to happen in the future.
New AI practices such as Generative Adversarial Networks (GANs) are well placed to analyse events like these. GANs can be used to repeatedly stress test the market by having two separate AIs work against one other. While one AI creates different scenarios, the other decides which information is real and which is false. This way, the AIs continually learn from one another over time, with one developing more complex data sets while the other improves its analysis techniques.
As such, GANs have the potential for businesses to evolve their processes and respond to shock market changes far more effectively. Not only can potential black swan events be tested with this model, but through this self-learning process, GANs can also develop their own hypothetical scenarios beyond human imagination.
With this level of predictive power, it is no surprise that GANs are seen by AI pioneers like Yann LeCun, the current Director of AI Research at Facebook, as the most interesting idea about machine learning to emerge in the last 10 years.
However, despite these advancements, humans still have a crucial role to play in the data management process. Manual oversight is needed to review results, ensure they are accurate, and remove any inconsistences or outliers. By following this process, the business can refine its AI algorithms over time to provide a more accurate image of the market and how it will respond to different events.
To support this, however, firms will need to democratise their data. Traditionally, a great deal of company information has only been available to a select few, but in order to utilise the potential of AI and GANs, this will need to change. Employees will need to have sight of data including trade history, investment yields and outcomes to better understand the AI’s decisions and amend it correctly. This will be especially important when it comes to managing a black swan event, since employees will need to have the knowledge, tools and ability to react in the right manner.
Despite AI’s advanced nature, this level of human involvement in the data management process is crucial. In every company, employees must have a solid understanding of how to sense check AI’s findings and ensure that data is consistent. This technology can act as the guide for workers and empower them to learn more about the business and the potential investment risks, but AI will ultimately deliver the greatest benefits through the combined efforts of both man and machine.
The global market is still very volatile, which means that black swan events will continue to cause assets and industries to shift in completely different directions. While black swans cannot be avoided completely, technology like AI can help businesses better prepare for these events by giving enhanced risk estimates of what may happen.