Artificial Intelligence


Minimal and Maximal Models in Reinforcement Learning

Authors: Dimiter Dobrev

Each test gives us one property which we will denote as test result. The extension of that property we will denote as the test property. This raises the question about the nature of that property. Can it be a property of the state of the world? The answer is both yes and no. For a random model of the world the answer is negative, but if we look at the maximal model of the world the answer would flip to positive. There can be various models of the world. The minimal model knows about the past and the future the indispensable minimum. Conversely, in the maximal model the world knows everything about the past and the future. If you threw a dice the maximal model would know which side will fall up and would even know what you will do. For example, it would know whether you will throw the dice at all.

Comments: 11 Pages.

Download: PDF

Submission history

[v1] 2018-08-25 15:00:52
[v2] 2018-09-17 09:47:30

Unique-IP document downloads: 40 times is a pre-print repository rather than a journal. Articles hosted may not yet have been verified by peer-review and should be treated as preliminary. In particular, anything that appears to include financial or legal advice or proposed medical treatments should be treated with due caution. will not be responsible for any consequences of actions that result from any form of use of any documents on this website.

Add your own feedback and questions here:
You are equally welcome to be positive or negative about any paper but please be polite. If you are being critical you must mention at least one specific error, otherwise your comment will be deleted as unhelpful.

comments powered by Disqus