Dr Florain Matthaeus Spiegel, co-founder at FinFabrik gives us a compelling insight into the rise of the Machines: an Introduction to AI and its application in Capital Markets at the 2017 Fintech Challenge here in Bangkok.
The rise of Machine learning with the ability to mimic our brain and neural networks is here and on the rise. Sensors are important to collect data - let's imagine all the data ever collected by mankind - 90% of all that data was collected by us and the data doubles every 2 years. In the future, the data will double every 24 hours. Data is like oxygen for AI. It needs Data to survive and learn. We now have more powerful machines and computers and they are becoming cheaper to create.
We have much better algorithms now between universities, startups and businesses.
AI is now everywhere. It is in our consumer products from Netflix, Siri to google maps all containing AI. The 3 main areas of AI combine Neurology, Technology and Biology. It's basically neural networks that try to mimic our brain.
There are different ways to do AI. The dominating approach comes from the different connections. Connectionists which use neuro networks to transmit Data.
How does this work?
Our brain has over 1 billion neurons. Each neuron has around 1000 of these synapsis connectors.
There is a signal running through our neurons called an electrochemical signal. It goes from one neuron to the next which we call 'Neuron firing.' This is where the connection occurs and the synapsis strengthens. What can we do with technology? We can try and replicate that on a very simple level. We will build a lot of layers into these networks which are called deep learning or the AI machine.
An Investment Management Renaissance Technology has been around since the 80s and has been developing AI. They develop algorithms and then let the machine make the investment decisions. They are extremely accurate and are very good at making investment decisions. We let it invest into cryptocurrencies.
The cryptocurrency is still very young and is like the wild west - we don't know enough yet. We need more regualtion in it. There is still a lot of opportunity in this market and the machines are the best at it for deciding and navigating the landscape. You can use a standard approach and you can also use a recuuring AI system. At some point we will limit the way AI speculates in this market as we as humans will want to take more control, we will do something that we did with High-frequency trading.
A machine is perfect for looking at data, making an analysis of it and then implementing a strategy. A machine is perfect at analysing. Efficiency is so accurate with machines. We should embrace the development of machines.
Full stack AI in capital markets such as chatbot, a special kind of AI is becoming a viable customer service channel combining layers of specialised AI to a full stack service platform. One thing we are excited about is that this platform can be offered to anyone. You can give me 5 dollars and I canIt creates financial inclusion - you don't need to give me millions anymore for wealth management. you can give me 5 dollars and I can offer the service to anyone with a fully automated platform at a very low cost.we will see challenges tho. It won't be plain sailing.
There is AI being used for cyber attacks and tracking these attacks. You can use AI for defending these attacks. Humans are doing this right now. Looking for computer locks and problems in the system, this is perfect for machines to do. There are a lot of companies in this space who are building solutions like this for cybersecurity. MIT is combing unsupervised learning - automated pattern detection which can for investment decisions to observe activity in the network but they also have the human side so the system is learning from the human - supervised learning.
MIT is combing unsupervised learning (Automated pattern detection) For investment decisions to observe activity in the network they also have the human side so the system is learning from the humans (supervised learning)
"AI will also have an impact on the economy with a massive increase in productivity.Instead of looking at the negative, AI will be like a colleague that is incredibly smart in one area and very dumb in another. We shouldn't feel threatened as it will be very helpful to do our jobs. Yes,the workforce will take a hit and it's something we need to understand that it will happen."
Goldman Sakks has replaced white collar jobs such as entry-level investment banking analyst which are high paid jobs actually and its already happening now. The real challenge is its happening very fast. If we look at the industrial revolution it happened over 30 - 40 years and society had time to adapt. This is happennig much faster than that. we will not have a lot of time to adapt to what AI will do to our workforce.
If we look at the industrial revolution it happened over 30 -40 years and society had time to adapt. This is happening much faster than that. We will not have a lot of time to adapt to what AI will do to our workforce. We will always be in control tho. AI will not take over everything. It's about continuous learning and will be able to move to the next level and be more creative and innovative that humans are made for while the machines do the jobs we no longer have to do.
To quote the founder of Alibaba - ' machines will be much smarter than we are but they will never have wisdom'. Lets embrace the revoluton including financial services that includes everyone.
Dr Florain Mattaeus Spiegl, co- founder at FinFabrik http://www.finfabrik.com/en/index.html
is a broadly experienced entrepreneur and Financial Technology practitioner, bringing transformational change to the organisations I work with. Focused execution is my leitmotif.
His work in FinTech has been centred around Digital Wealth Management and Capital Markets software solutions that unlock unique client experiences by leveraging technology. I bring my experience to established companies and startups as a speaker, advisor and mentor, along with running FinFabrik, a FinTech Solutions Builder.