Everyday it becomes more and more often to listen to stories that used to shape our fictional futuristic worlds: self-driving cars, that improve their driving skills with experience, refrigerators that notice when you are running low on something, and order it online for you; financial apps that let you know when you are getting out of your budget. All of these products are the result of machine learning.
Machine learning is the subdivision of Artificial Intelligence that emulates the human brain, and gives the machines the capacity of learning through the analysis of large volumes of information.
We all have heard about the machines that beat professional Chess or Go players by studying human movements and improving them, but the AlphaGo Zero taught itself to play Go only by analyzing only the rules of the game. It then discovered all the human gameplays and then developed better, never seen strategies.
Thanks to these qualities, machine learning has started to be applied to finances, where the possible applications are huge, just think about it: the analytical capabilities a machine has, how they can detect patterns, and can predict them to some degree; all the desirable qualities of a financial advisor, plus the unlimited learning power machines posses.
And even though this subdivision of Artificial Intelligence is still on its infancy when it comes to Finance, we can start seeing them across the Fintech sector; Datamir uses an algorithm to scan tweets and single out important events, even before they’re reported, Kensho analyzes all kinds of documents to answer to 65 million different questions about where the market is going; Acorns not only rounds up your spare change and invest it, but lets you know when you are getting out off your regular spending patterns.
We can find other applications of machine learning in Finance on areas such as risk management, compliance and fraud prevention. Intelligent Voice, a British company, markets its voice transcription tool to financial institution to audit their telephone operators; Monzo, also British, created a model that prevents fraudulent transactions to be completed, reducing the rate of fraud from .85% to less than .1% in less than a year.
We might not be close to have an A.I. managing our hedge funds, tailoring our investment portfolio to our personal profiles, guarantying the best returns without the risks that come with the human factor of financial speculation.
But we are definitely not far.
The power that machine learning has to go beyond the brightest most capable human minds can make anyone nervous. What kind of future is waiting for us? To a future where machines become aware of themselves and our place as the dominating species? To one where machines never transcend their status as a human tool?
But while the future reach us, would you trust your finances to a machine?