Who is James Harris Simons or the “Quant King”? He is a self-made billionaire ranked by the Forbes, records Nickolas (2016). For more specifically he is a scholar, company owner, military code breaker. After studying maths for a long time, he worked for the government as a code breaker. After that, he decided to pursue a career in finance and started a company with a friend from the same “code-breaking place”. With computer models and algorithms, his company became very successful. And he used machine learning as his tool to achieve this success Simons said in 2015. Like machine learning, AI is in many areas; from stock market management to traffic light synchronization; from managing Facebook timeline to chess that people play against the computer. Each usage area has its subset of AI; deep learning, machine learning, neural networks, etc. And of course, finance is in one of these areas. Artificial intelligence, as a tool, has benefits in finance, namely; credit evaluation, portfolio management, and financial prediction and planning. Although AI seems to cause unemployment in many areas, its development should not be prevented, because it also helps to create new business opportunities for humans.
Usage of the AI and its applications are beneficial in credit evaluation thanks to the specialties of AI which can be listed as having unlimited capacity, being fast and accurate when compared to human beings. As a subset of AI, neural networks (NN or ANN) have those qualifications. NN is a computer algorithm that learns directly from the given data (Hammerstrom, 1993). One of the beneficial usages of AI in finance is credit evaluation, which is simply a typical classification problem that is deciding the value of a person or corporate entity according to predefined categories (Bahramirzaee, 2010). With the ability to mimic, NN can classify the considerable amount of customer data like an expert, Bahramirzaee describes in his article: A comparative survey of artificial intelligence applications in finance; artificial neural networks, expert systems, and hybrid intelligent systems in 2010. Using NN is beneficial in credit evaluation, because it is better than the old methods for credit evaluation, as we see from Lloyds Bowmaker Motor Finance Company example. They used ANN for their car’s financial decisions based on credit scoring. And got the result with %10 more success. As is shown in credit evaluation, another usage of NN as a part of AI in finance is portfolio management.
AI is important as it contributes to portfolio management where it eliminates the risks in the finance area. The uncertainty of the economic environment and the diversity of information can be balanced with making accurate predictions thanks to the algorithms of AI. As one of the challenges of portfolio management, portfolio selection is finding the most appropriate way to deposit a certain amount of money into several investment opportunities, and each possible way of making these investments are called a portfolio (Fernandez and Gomez, 2005). ANN can make predictions by correlating and considering lots of independent variables. Such a specialty can be really helpful for managers with deciding on where to invest. Especially for portfolio managers, giving decisions can be very tricky since they have to choose between several options. Therefore, ANN can be used for smarter decisions. Bahramirzaee (2010) reports that NN is used for portfolio management which is tested for 21 different markets of G7 countries and concludes outperforming results over the traditional management methods. This was a good proof of the benefits of NN in the field of portfolio management.
Finally, AI can be used as a useful tool in financial prediction and planning since it has unique abilities as being suitable and capable in financial applications. As Bahramirzaee (2010) classified the financial market as not stable so that its interactions and requirements are difficult for people to understand. Hopefully, NN has the nature which is suitable for financial predictions. According to Bahramirzaee (2010), NN has significant success among financial prediction and planning. Referring to him, NNs properties are unique and suitable for this purpose because they are naturally numerical, require no assumptions and models as inputs and are capable of updating data. To give an example to this, Celik and Karatepe’s research can be used. In their study, NNs work on evaluating and predicting the banking crises studied. And they see that NN managed to evaluate and predict the banking crisis by finding solutions for semi-structural and non-structural problems (Bahramirzaee, 2010). Looking at the capabilities of AI, their success in finance, handling complex problems successfully, one might think that there will be no need for human work in the future.
Even though AI seems to cause unemployment, it does not. Researchers show that the development of AI could create new jobs. According to Wilson (2017), large companies who are already using AI technology observed that the field was also creating new jobs that require humans with skills and training for specific tasks of AI development. So, one might say that investing in AI, not preventing its development might not actually cause unemployment. Regarding this, we should not prevent the development of AI. Humans are tool-making animals, AI is one of them, and as we did in the past, we are heading into the future by inventing tools.
To conclude, AI, especially Neural Networks, is a very beneficial tool for finance. Finance as a complex, unpredictable, dynamic subject, has features called credit evaluation, portfolio management, and financial prediction and planning. Compared to humans, AIs (NNs) are better at this task by nature. Seeing these qualifications, we might feel useless and anxious about the future, but it is not necessary. We should not prevent its development and be more optimistic about the future.
References
Bahrammirzaee, A. (2010). A comparative survey of artificial intelligence applications in finance: artificial neural networks, expert system and hybrid intelligent systems. Neural Computing and Applications, 19(8), 1165-1195.
Fernández, A., & Gómez, S. (2007). Portfolio selection using neural networks. Computers & Operations Research, 34(4), 1177-1191.
Hammerstrom, D. (1993). Neural networks at work. IEEE spectrum, 30(6), 26-32.
Nickolas, S. (2016, March 6). Jim Simons’ Success Story: Net Worth, Education & Top Quotes. Retrieved July 23, 2019, from https://www.investopedia.com/articles/investing/030516/jim-simons-success-story-net-worth-education-top-quotes.asp
Numberphile. (2015, May 13). Billionaire Mathematician – Numberphile. Retrieved from https://www.youtube.com/watch?v=gjVDqfUhXOY&t=927s