Artificial Intelligence In Finance: Half A Century Journey From Science Fiction To Cutting-edge Financial Advisory

HI vs AI

[Source: © TROYZENTRO | DREAMSTIME.COM]

Human VS Artificial Intelligence In Finance – Shall We Compete Or Collaborate?

It is a widely known phenomenon that a human brain cannot depict on a painting something that has not been presented in nature wholly or at least partially. As such, Artificial Intelligence (AI) stood out on the edge of human imagination and philosophical discussions for a long time being something which is presented inside humans and nowhere else. Nowadays, artificial intelligence is already part of our lives and it improved our lifestyles, although we sometimes take it for granted and do not want to go into details of what is “under the hood” of those systems. As such, our online shopping experience became much faster due to constantly improving recommendations or the way we listen to internet radio stations via streamline services like Google Music or Spotify.

AI has been part of scientific research since a handful of computer scientists rallied around the term at the Dartmouth Conferences in 1956 and birthed the field of AI. Back in that summer of ’56 conference the dream of those AI pioneers was to construct complex machines that could have the same features as human intelligence. This is the concept we think of as “General AI”, i.e. machines that have all our senses and even supernatural ones, all our reason, and think just like we do. Even though the computer technology exponentially advanced over the last 30 years, General AI machines have remained in the movies (take Terminator, for instance) and science fiction novels as we can’t technologically pull it off so far.

The progress which computer science made so far gives us the right to call the most advanced AI based technologies as “Narrow AI.”, i.e. technologies that are able to perform specific tasks as well as, or better than humans can. Examples of narrow AI are things such as image classification on a service like Pinterest or high speed advanced analysis of Big Data emerged from financial markets, taking into account incredible number of parameters and factors affecting companies’ performance over time.

Machine Learning As A First Step To Achieve Artificial Intelligence

Picture 1: AI, Machine Learning and Deep Learning Hierarchy

Picture 1: AI, Machine Learning and Deep Learning hierarchy

The most common way to think about the relationship between AI, Machine Learning and Deep Learning represent them by depicting AI as being the largest concept that emerged first, then machine learning, which went later, and finally deep learning, being driver of the today’s AI explosion going inside both of the previously mentioned concepts.

Machine Learning at its most basic is the practice of using algorithms to gather data, learn from it, and then make a determination or prediction about something in the real world. So rather than hand-coding software routines with a specific set of instructions to accomplish a particular task, the machine is actually able to “train” itself using large amounts of data and algorithms that give it the ability to learn how to perform the task.

Machine learning came directly from minds of the early AI crowd, and the algorithmic approaches over the years included decision tree learning, inductive logic programming, clustering, reinforcement learning, and Bayesian networks among others. As we know, none achieved the ultimate goal of General AI, and even Narrow AI was mostly out of reach with early machine learning approaches.

Machine Learning Evolution – Deep Learning & Artificial Neural Networks

Another algorithmic approach from the early machine-learning crowd is to use Artificial Neural Networks. Neural Networks are inspired by our understanding of human brains and its neurons empowering us to solve most complex tasks. Unlike an actual brain where any neuron can reach to any other neuron within a certain physical distance, an artificial neural network may have discrete layers, connections, and directions of data propagation.

For example, you may take stock performance data, chop it up into a bunch of time periods that are inputted into the first layer of the neural network. In the first layer individual neurons, then passes the data to a second layer. The second layer of neurons does its task, and so on, until the final layer and the final output is produced. Each neuron assigns a weight to its input based on the correctness or incorrectness relative to the task being performed, i.e. the machine learning process is performed.  The final output is then determined by the total of those weightings.

Pic 2: Deep Learning and ANN

Picture 2: Example of artificial neural network with 3 layers, machine learning process occurs on each of the layers. The functions above are general output functions.

So think of an Apple Inc. stock price example. Diverse financial and other factors affect the stock performance over time and data describing could be “chopped up” and “examined” by the neurons. The neural network’s task is to conclude whether a specific stock may show bullish or bearish behavior and how certain such prediction could be. So in our example the system might take as input lots of stock market data in respect of the stock over the period of 15 years back until today and produce forecast output for those stocks in terms of signal strength and predictability. In order to do that the system will process the inputs layer by layer and weight the respective inputs based on its previous success-to-failure history. Ultimately, it comes up with a “probability vector,” really a highly educated guess, based on the weighting and a sample prediction for the stock may be bullish behavior represented by a positive signal with certain numeric value for 1 year forecast horizon with 59% predictability.

I Know First Algorithm Serving The Investor – Artificial Intelligence System That Is More Than Real

Artificial Intelligence and Machine Learning are two cornerstones which provide the basis for the I Know First predictive algorithm to produce forecasts on a daily basis. The architecture of the algorithm is inspired by the latest trends in AI and Artificial Neural Networks supported by Genetic Algorithms approach incorporated to it. The last concept provides the algorithm with means to create, modify, and delete relationships between different financial assets being considered within the forecasting process. Literally, it constantly proposes “theories”, testing them on years of market data, then validating them on the most recent data, which prevents over-fitting. If an input does not improve the model, it is “rejected” and another input can be substituted. Since the algorithm learns from its previous forecasts and is continuously adapting the relationships, it adapts quickly to changing market situations.

Pic 3: I Know First value Proposition

Picture 3: I Know First value proposition

In financial terms the I Know First Market Prediction System models and predicts the flow of money between the markets. It separates the predictable information from any “random noise”. It then creates a model that projects the future trend for a given market in the multidimensional space of other markets. Therefore, the system’s prediction output adds significant value to investment decisions to be taken, as they provide numerical insight on relative attractiveness of investment into a specific stock in terms of its expected behavior trend and corresponding predictability.