Home / In evidenza / INVESTMENT FUNDS AND NEURAL NETWORKS: A CHALLENGE TO MAKE MACHINES PROTAGONISTS

INVESTMENT FUNDS AND NEURAL NETWORKS: A CHALLENGE TO MAKE MACHINES PROTAGONISTS

Man has to deal with ‘fear and greed, intellectual constraint and fatigue’, whereas a machine is ‘agnostic, tireless’ and has ‘no bias’ in decision making – Standard Life

ddArtificial intelligence as books say, “it ‘s a branch of its own“. More than a school of thought is a set of techniques that aim to simulate the process of human reasoning (expert systems), the biological functioning of the brain (neural networks) or the evolution of living species (genetic algorithms). Irrespective, in itself, by the principles of the technical thought and the fundamental as it.

Among these techniques this article wants to focus its attention on neural networks, and the meaning is that an interesting field of research and development in science that studies the neural networks, is that of financial applications, and in particular the construction of systems for anticipating the future performance of financial instruments such as equities, interest rates, stock index, currencies.

An artificial neural network is a computer tool that mimics the operation of a biological brain to store and to use the information received. As well as the brain in nature stores the information in a distributed manner, in a network composed of countless neurons connected to each other by synapses more or less equipped with electrical conductivity, an artificial neural network consists of many processing units that, unlike what happens in traditional software programs, acting in parallel and not in series, connected to one another by connections.

So neural networks, unlike the previous statistical forecasting systems, which were based on rules and on financial models defined a priori, model the mathematical system which must then simulate the performance of the stock in question, finding an almost natural way, connections and correlations on the set of past data, that is, by analyzing historical data on a number of variables you want to look for collaboration. These relationships existing on past data, compared with the subsequent behavior of the title, are found through a mathematical process that ties together the importance and influence of each variable in order to generate a final prediction as correctly as possible.
Currently it seems that the application of neural networks in the field of financial forecasting is giving good results, given that the managers of the most important World investment funds argue that in a not too distant future adaptive tools for financial forecasting equipment will become mandatory for all professionals.

Nowadays in fact more than half of financial transactions in the world are automated, made by machines suitable to perform several thousands of operations per second. Even more traders is artificial intelligence that speculates more than any other.

Due to the high results that AI can give, investments firms argue that many jobs performed by fund managers could be replaced by machines: “Using artificial intelligence applications have enhanced our understanding and analysis of financial market behavior, adding to the range of predictive tools”.

To support what has been said before, Standard Life also asserts that “Man” has to deal with “fear and greed, intellectual constraint and fatigue”, whereas a machine is “agnostic, tireless” and has “no bias” in decision making.elle

Nonetheless replace at all the figure of men is not yet possible, since computers are able at most to varying extents to add value to quantitative inputs, and human thinking it is better suited to the qualitative side; as this firm underlines, “investment approaches generally contain both qualitative and quantitative elements”.

The characteristics of human intelligence, that AI has not yet reached, are in fact creativity, judgment, and intuition, all qualitative features.
The strength of a processing system instead is given by speed, accuracy and attention to detail.
It is well understood so that being able to create a system that has both the human capacity that the ability of a processing system can obtain a product usable in any type of process and can give a certain guarantee of result.

Conclusions

The great novelty brought by the techniques of artificial intelligence is the ability to use the machine in processes not fully modelisable. This is the great advantage that you get compared to using the traditional techniques of software applicable only in structured processes, which are known a priori the activities to be performed and the sequence in which they should be performed.
Of course the current artificial intelligence techniques are not yet able to replace humans, even if they have in this case enhanced really the understanding and analysis of financial market behaviour, but gradually are becoming a valuable support to the decision maker in increasing human situations. So the big challenge for the future is to be able to amplify the capacity of the intelligence of the computer through a link with human intelligence.

 

 Useful Links:

http://en.wikipedia.org/wiki/Artificial_intelligence

https://mitpress.mit.edu/books/ai-business

About Vanessa Lumini

The genius build the world, the clever turn around, and the stupid think that the world revolves around them .. Laureata e appassionata in Informatica, contabile amministrativa di professione da 5 anni, con la voglia di diffondere le mie passioni in tutte le lingue a me possibili.

Leave a Reply

Your email address will not be published. Required fields are marked *

*

Scroll To Top