Friday 3 March 2017

Imperfect information

In the actual incarnation of the fractal AI, we need to supply it with its exact state, all the interactions with environment (the simulation) and the potential function, hand crafted, to be followed. This is know as having "perfect information".

Having perfect information of any system is just not feasible, so my models are not usable in real environments, with real drones, moving real motors, as all of them are unknow for us and we will have just some sensor's outputs as our information.

This week I have visited the Cognitive Sciencies research team at Zaragoza University as a guest for a short but intense seminary about my fractal inteligence algoritms in an effort to team with them here and there, but they really emphasized on the sensorial approach -imperfect information case- in order to make our works compatible.



So I have been thinking on how the actual fractal AI can be reduced to having only a number of sensor of the world around -and its on internal state- and some motors or actuators to work with. May be I will have it adapted in a couple of months, as I already have an easy way to properly do it, but it doesn't fit on the margins of this book ;)

The idea is basically to mimic how a baby learns how to use his muscles to move around just form the information it gets from some sensor, in a continuous proccess of trial-and-error as you move around randomly, learning so, with time, the correct/optimal behaviour just emerges.

I will be posting news about it as I develope the idea into real code and make some new videos of new-born rockets with only some distance sensors around it as they learn to fly themselves.

BTW, there will be a long video of the first session of the seminary -the other two sessions were more informal and were not recorded- where I explain the fractal inteligence at deep. I will also publish the presentation  (I need to translate it to english first) along with a big doc I have crafted with all the theory and pseudo-code I use, so you could produce your own fractal AIs in the same level -or even higher, as it includes the fractal memory schema, something I have not showed on videos on this blog- than the ones you have seen in the latest videos. Also the python code we use to work on some OpenAI problems will be released.

So keep tuned!

2 comments:

  1. I think this will be a big challenge. The algorithm would basically need to learn to model arbitrary data time series coming from arbitrary sensors, so it can 'predict the future' for the AI algorithm. I don't know of any algorithm that can do this; even neural networks are extremely limited in this respect... Wish you the best of luck, excited to see what you come up with.

    ReplyDelete
    Replies
    1. Hi Juan, actually it is a big challenge I never dream on working at, and may be my inital ideas are not correct "out-of-the-box", but I think I can get something out of it.

      My first step is to learn from noise signals, so data time serie needs to be converted into some kind of form that, at the same time is good at representing the system state and predicting the new signals in some basic form, and also form a "pattern" of what happened to the signal prior to getting into this state.

      I am playing with a form of "Taylor serie" to use of the "ststae" of my signal, so this state both represent the actual state and its tendencies, along with a pattern formed by the last N received signals.

      Once this "sensorial state" or "Taylor state" is done, I will start the "fractal part" of mixing several of those states into a predictive memory.

      So this is the basic ideas, intuition tells me it could work, but actual code is not always doing what one spects, so time wil tell!

      Delete