Google's Deep Mind "Alpha Zero" - amazing AI concept prooved in GO and chess - page 3
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
Alain this is a misconception.
The current state of the technology is that they do learn to play by themselves they re-write their own code.
I know that. Where is the misconception ? Are you thinking this AI will start to create code to solve all kind of problems by itself ? An AI is not able to get out it's algorithm by itself.
That's what a human intelligence is doing, but 1 human is just too limited in speed and time.
I know that. Where is the misconception ? Are you thinking this AI will start to create code to solve all kind of problems by itself ? An AI is not able to get out it's algorithm by itself.
That's what a human intelligence is doing, but 1 human is just too limited in speed and time.
But even without it, the uses are enourmous. It can learn very quickly things that humanity needs hundreds of years.
Medicine formulas, science etc.. All of this, are exactly the strong side of this kind of machine learning.
It is not a theoretical debate, if it's really a human inteligence or not.
It is about facts, and the facts are:
AI today is an algorithm of course, but the writer of the algorithm (human) does not know the final outcome of it, and even in succeeding (the AI plays better) he does not know why the AI played a move (for example, if it plays chess, you as the programmer can't tell why the AI chose certain move, and you can not influence it after it learned).
This makes a seperation between non-AI programming - where you need to tell the program the rules, and modern AI- where the algorithm is able to learn the connections, even if you yourself don't know those connections.
Call it what you want, it does not matter. I see it as valuable in trading as well.
Don't be irritated please. I agree with you.
I was discussing a step further, I was probably not clear. That's enough for this site, don't start to talk about intelligence here.
I know that. Where is the misconception ? Are you thinking this AI will start to create code to solve all kind of problems by itself ? An AI is not able to get out it's algorithm by itself.
That's what a human intelligence is doing, but 1 human is just too limited in speed and time.
This is exactly what i mean, it is able to get out of it's algorithm, because it is able to re-write itself.
I don't think you understand where it's currently at.
It can and will climb out of it's own box because we have programmed it to allow that, and that is different from fixed or static code.
These algorithms are truly self learning and they are a threat because of the speed at which it happens.
We are not certain that we can contain and control the technology when it is allowed to evolve.
I will give you a very serious example, give enough time, the AI will learn, by itself, that it can communicate over radio waves, by moving electrons over wires and metals.
This is exactly what i mean, it is able to get out of it's algorithm, because it is able to re-write itself.
I don't think you understand where it's currently at.
It can and will climb out of it's own box because we have programmed it to allow that, and that is different from fixed or static code.
These algorithms are truly self learning and they are a threat because of the speed at which it happens.
We are not certain that we can contain and control the technology when it is allowed to evolve.
I will give you a very serious example, give enough time, the AI will learn, by itself, that it can communicate over radio waves, by moving electrons over wires and metals.
Yeah for sure it's me who don't understand. Please give me a reference that explain or show that an AI algorithm can, right now, get out of its algorithm. Science fiction movies are not accepted.
EDIT: Just thought maybe there is a misunderstand about I a mean by get out of it's algorithm. So don't lost your time trying to convince me. People build systems and they can't explain how these systems are doing what they are doing, and they start to be afraid, and they start to scare others about AI threats. Finally maybe that's true the AI machine are more intelligent than humans.
https://www.youtube.com/watch?v=gn4nRCC9TwQ
Here they allowed one to learn how to walk just like a human being.
Not really sci fi a la startrek but it is possible to write self learning software.
To simplify the way the algorithms work, they are given a way to measure for themselves and decide between one behavior and the other, which one scores better. And they constantly strive to perform better, by trying different behaviours randomly (in case of unsupervised learning).
This is why they are unpredectible, because this type of reinforcement learning is controlled by the AI itself, making random steps and then decides which one takes it forward and upward better then the other (Methaporically, and also practically in the case of the walking and running robot that Marco showed). In case of chess it can be who wins and in how many moves etc.
Google’s DeepMind today shared the results of research and experiments in which multiple AI systems were trained to play Capture the Flag on Quake III Arena, a multiplayer first-person shooter game. An AI trained in the process is now better than most human players in the game, regardless of whether it’s playing with a human or machine teammate.
The AI, named For the Win (FTW), played nearly 450,000 games of Quake III Arena to gain its dominance over human players and establish its understanding of how to effectively work with other machines and humans. DeepMind refers to the practice of training multiple independently operating agents to take collective action as multi-agent learning.
“We train agents that learn and act as individuals, but which must be able to play on teams with and against any other agents, artificial or human,” the company said today in a blog post. “From a multi-agent perspective, [Capture the Flag] requires players to both successfully cooperate with their teammates as well as compete with the opposing team, while remaining robust to any playing style they might encounter.”
I have been studying the white paper of alpha zero and it's algorithm and trying to figure out how to implement it into MQL5.
So far I found very little MQL5 resources as how to create a similar algorithm which can converge over time like "Alpha zero" by self learning and self training which can work in forex market.
In my understanding, if we can combine Q learning along with random decision forest along with automatic optimisation , then it can be possible. But I am not an expert programmer to combine all these areas by myself.
If any good programmers are there who are still interested to develop a similar algorithm like "Alpha zero", then we can continue the discussion further in this thread regarding the various ways of implementation and later on converting the algorithm to MQL5 code.