# Re: [Cleveland-AI-ML-support-group] Study Progress

 From: Joe Sent on: Friday, September 2, 2011 10:59 AM
Interesting point. Given the way the real world is, it would seem that that approach would often be the right one. Another approach that seems related to that is curve fitting, such as with Newton's method for squares. I'm not sure how curve fitting would be classified. Is approximating with an infinite series close enough to deterministic to be called deterministic? It looks like the class schedule is only going to cover chapter 3 for search, which is a bit disappointing for me because I'm into that area. Ch 4 looks like it gets into something like curve fitting, and it also has a section referring to search and none determinism. I guess that will be semester 2 material. It looks that semester 1 will cover the basics of each area, and semester 2 will fill them in.

This brings up this other thought, that the 5 basic types of agents (reflex,reflex model,goal,utility,learning) may all be used by one agent system at different levels of world state resolution. For a robot taxi a goal based agent may use probabilities to calculate stopping distance, and when the taxi is closer than stopping distance to the next car with no cars behind it, a reflex agent triggers the brakes on that signal. I think the book may have had a line in it somewhere to that effect, does anybody remember that line?

New members may be interested in the growing flash card set as a way to catch up if needed, it's linked on the wiki: http://freedombluesky.com/aiml

Joe

--- On Fri, 9/2/11, Timmy Wilson <[address removed]> wrote:

Subject: Re: [Cleveland-AI-ML-support-group] Study Progress
Date: Friday, September 2, 2011, 6:56 AM

> The problem is that the number of states quickly
> grows beyond your ability to analyze them and
> necessitates a different approach.

I've been hung up on determinism lately

This is from the wikipedia entry on determinism -- http://en.wikipedia.org/wiki/Determinism:

Many mathematical models of physical systems are deterministic. This is true of most models involving differential equations (notably, those measuring rate of change over time). Mathematical models that are not deterministic because they involve randomness are called stochastic. Because of sensitive dependence on initial conditions, some deterministic models may appear to behave non-deterministically; in such cases, a deterministic interpretation of the model may not be useful due to numerical instability and a finite amount of precision in measurement. Such considerations can motivate the consideration of a stochastic model even though the underlying system is governed by deterministic equations.

The important piece:

Such considerations can motivate the consideration of a stochastic model even though the underlying system is governed by deterministic equations.

On Thu, Sep 1, 2011 at 11:04 PM, Joe wrote:
 Can't help but want to correct this: "In chess that number depends most of all on the 'material' score of the pieces."Should instead be:"In chess that number depends most of all on first the opposing king being checkmated or not  and second the 'material' score of the agents own pieces. The checkmated king scoring 1000."

--
This message was sent by Joe ([address removed]) from Cleveland AI + ML support group.

Meetup, PO Box 4668 #37895 New York, New York[masked] | [address removed]

--
This message was sent by Timmy Wilson ([address removed]) from Cleveland AI + ML support group.

Meetup, PO Box 4668 #37895 New York, New York[masked] | [address removed]

### Cleveland, OH

Founded Aug 19, 2011

#### Organizers:

• ##### Bizdom Cleveland

Support, lovely space in tower city, lunch..

#### People in this Meetup are also in:

• ##### The Cleveland Java Meetup Group

701 Java Developers/Architects

2,488 Gamers

• ##### The Cleveland Professional 20/30 Club

1,473 Young Professionals