While I believe that machine learning will change the world in finite reasoning domains (judges, lawyers, driving, doctors, teachers, etc...all whom will lose their jobs soon to these machines) from my point of view there is one aspect of it that leaves it wanting in infinite adaptable domains. Imagine that a machine was trying to prove the Riemann Hypotheses, and it came up with a proof that looked like this: (0111,333,22,111199977752314159, 6947,...900 trillion numbers later, QED) Every machine in the world, using whatever theory was used to construct such a sequence, verified that the Riemann Hypotheses was indeed proved. What good would it do other than the proof itself? The problem is that computers, at least as we understand them today, compile down. Human beings compile up. What I mean is that computers just build an incomprehensible assembly language of correlations with finite nodes to make inferences. Whereas human beings compile up to general emerging principles . It is hard to state, but what I think I am getting at is that ML and computers in general are trying to prove theorems, while human beings are going up and down the reasoning stack simultaneously, and most importantly, intuiting General Theories that don't exist at all in any of its finite systems of current understanding to emergent general principles of reasoning. Said in another way, human beings, particularly artists of which I include mathematicians, are building higher level Godel models/languages to make statements in lower level languages vastly easier to understand. Imagine the following situation in applying a similar ML technique to trading. Say you build a trading system using the same technology as used by Googles DeepMind, or something even more advanced like the system above that goes on to prove the RH. Imagine that you spend a huge effort to make a system that seems to work, from the point of view of profitability, but you don't understand a single reason as to how it is making decisions. Still, you are brave and you turn it on live. The system does great for a couple of months, then it is breaking even, then it starts to lose. What do you do? So, the problem with ML imo particularly as it applies to trading, is that since the human being cannot understand the output of the machine, and since the ultimate node in a system is the human being deciding whether the system was viable or not after some adverse run, it leaves us in an anxious state both while running it and while deciding to take it off line. That is not a trading foundation to build on.
Note that I am not saying that automation is wrong. Hardly. I am saying that ML is a poor edifice on which to build trading systems because they are, at least currently, black boxes even to the people that decide whether to turn them on/off.
On the other hand: Facebook robot is shut down after it ‘invented its own language’ Scientists have switched off Facebook’s robot because it has reportedly created its own language. Researchers noticed that Artificial Intelligence they had created has started to make up their own code words. Initially it looks like absolute gibberish, but it became clear that the machines – nicknamed Bob and Alice – were actually communicating with one another. In a Frankenstein twist on their research, the robot abruptly stopped using English and could only be understood by other AI. It comes after other AI developed in a similar pattern elsewhere, like when Google Translate invented its own language. The so-called ‘neutral network’ started translating phrases easier using it’s own language, according to the New Scientist. Read more:http://metro.co.uk/2017/07/31/faceb...ented-its-own-language-6818204/#ixzz4qE3DNCR7
How can AI do any better that identifying and riding a trend, which (some) humans have doing since the first markets opened?
I don't have a lot of experience with ML or AI. That being said, the new backtesting platform I am building in Golang does use some supervised learning techniques. My goal is to be able to have an hypothesis which will include which features may represent a pattern, and then the system will go through each point in the time series, get the state of those features, and then see if those collection of features represents a trade-able pattern. My goal is to automate, beyond the initial idea and setup, how a human would find an edge. A very simplistic example would be the state of the last four closes. Of course I would not expect that to be a trade-able pattern, but that is the approach I am taking. The challenge is I may be interested in a collection of features and want to run each combination of them. That can result in a high number of iterations but I have found techniques for reducing the number of combinations I am really interested in. I will post more as my project moves along. fan27
Let's call it ML since AI can mean a bunch of things, or nothing at all. Linear regression is ML, so very doubtful any big institution are not using some form of ML. From Wikipedia: Machine learning is the subfield of computer science that, according to Arthur Samuel, gives "computers the ability to learn without being explicitly programmed." Samuel, an American pioneer in the field of computer gaming and artificial intelligence, coined the term "machine learning" in 1959 while at IBM. So while there's lots of new hype surrounding ML, it's nothing new. There are recent breakthroughs in various practical applications and hardware implementations that should concern anyone who consider themselves superior to machines. Cutting it short: Humans will use machines to get an edge over you, and then blame you before the machines finds a solution for that as well. Such socio-economic violence is already disenfranchizing millions of people, and why the political climate is the way it is today. Still there are humans designing these things, and they need to know how and why the program should converge to preferable solutions, so it's not all dark magic.
AI and ML lastest buzz words both utter BS both so far off, we need a gigantic leap before either are real. Most are simple random number based decision making, thats guessing at best not intelligence. Bob and Alice failed, they both started saying Bla Bla Bla basically but in other text, there not intelligent they have nothing to talk about.
The movement in the markets is not random but the way humans cause price changes is beyond any man made algorithm. It has no rational logic to it and therefore will flummox both man and machine.