Term Project?

[Ali Helmy]

Being in the AI Society means that you have an 'extra' enthusiasm about AI... so what do you suggest could be a good Project for this term's AI-Course?

Remember, it includes 25 people, so it must hold to the magnitude

PS: is anyone still here? i seem to be talking to myself most of the time

--
Cheers,
     Seth
     <A. Helmy>
==================
  One Life... LIVE It
==================

[Khaled El Ghayesh]

Well, I don't like the idea of doing a 'Graphics' project for the 'AI' course. It is just that we need to concentrate on the course itself to grasp the maximum benefit from the project, not to waste our time on other irrelevant stuff (although it is important to know how to work in Graphics properly of course !).

I reckon it's better to make it something more logically sophisticated and less interface-centered. What comes up to mind is doing something related to words and language. That will make sure we really concentrate on the real gig not other stuff irrelevant to the course.

Well, I was just writing this to start brainstorming for ideas, but what came up to my mind was maybe to make a chatting BOT. An agent that tries to immitate normal intelligent human conversation. I didn't think much into how to do it and the details of its implementation but I guess we can come up to some interesting stuff.

Please say your feedback.

Khaled

---

From: "Ali Helmy" <alih...@gmail.com>
Reply-To: ai...@googlegroups.com
To: AISOC <ai...@googlegroups.com>
Subject: [AIS] Term Project?
Date: Mon, 9 Oct 2006 23:18:23 +0200

[Ali Helmy]

On 10/10/06, Khaled El Ghayesh <kgha...@hotmail.com> wrote:

Well, I don't like the idea of doing a 'Graphics' project for the 'AI' course. It is just that we need to concentrate on the course itself to grasp the maximum benefit from the project, not to waste our time on other irrelevant stuff (although it is important to know how to work in Graphics properly of course !).

I agree completely... if we go with this, u will have atleast one team [3 to 4 people] working on the graphics interface, and another team working on the networking... that means uve got around 8 people per group, making a total of 32 people [32%] who dont really get to work on the project...

I reckon it's better to make it something more logically sophisticated and less interface-centered. What comes up to mind is doing something related to words and language. That will make sure we really concentrate on the real gig not other stuff irrelevant to the course.

Well, I was just writing this to start brainstorming for ideas, but what came up to my mind was maybe to make a chatting BOT. An agent that tries to immitate normal intelligent human conversation. I didn't think much into how to do it and the details of its implementation but I guess we can come up to some interesting stuff.

well, id LOVE that since i love everything to do with NLP, but again... that really cant be cut down into a form of a competition, is it? how can u judge who had the best conversation? and you know, the AI projects competition is something that goes on EVERY year, and in my opinion is a good thing... just, the football thing is HORRIBLE

Please say your feedback.

Khaled

[Khaled El Ghayesh]

Yeah sure football is crap for an AI project. I am sure we will end up with all algorithms ready and functioning and everybody waiting for a mysterious bug in the graphics engine, so the whole thing will be useless, which is really ridiculous.

How we are gonna judge conversations ?? hmm that's a good point dude. I don't think there has to be really direct competition between different agents. We can have a specific standard from which to compare the intelligency of the conversations that took place. I mean look at sports, there are many sports where there is no direct competition between competitiors. Actually there is no contact between them at all. Everyone just does his/her part and the jury gives their mark.

What I wanted to say is that it is not about competition as much as it is about how to build a proper "working!!" project that does something useful.

 

---

From: "Ali Helmy" <alih...@gmail.com>
Reply-To: ai...@googlegroups.com

To: ai...@googlegroups.com
Subject: [AIS] Re: Term Project?
Date: Tue, 10 Oct 2006 01:10:13 +0200

[Ali Helmy]

On 10/10/06, Khaled El Ghayesh <kgha...@hotmail.com> wrote:

Yeah sure football is crap for an AI project. I am sure we will end up with all algorithms ready and functioning and everybody waiting for a mysterious bug in the graphics engine, so the whole thing will be useless, which is really ridiculous.

I can see it already

How we are gonna judge conversations ?? hmm that's a good point dude. I don't think there has to be really direct competition between different agents. We can have a specific standard from which to compare the intelligency of the conversations that took place. I mean look at sports, there are many sports where there is no direct competition between competitiors. Actually there is no contact between them at all. Everyone just does his/her part and the jury gives their mark.

I know, just like diving or even WRC ... and in that sense, how about the infamous TuringTest for judging the quality? we can even simulate it real-time with volunteers from Dr.s, TAs or even students from outside the class

What I wanted to say is that it is not about competition as much as it is about how to build a proper "working!!" project that does something useful.

You misunderstood me... i totally agree with you, im just saying that someone *cough* the doctor *cough* *cough* might not agree... bas!

[Sherif Elian]

Sorry i was and still little busy , but I am reading All mails , warak ya rayes :D

 

Anyway , I totally agree on Estimation , it is smart game, not so complicated , not so easy ,

Also there is no “The Right Play” , there is “A Good Play” and “ A Better Play” what I mean here is that in connect4 for example you can make sure your opponent will not connect 4 or even make a double sided three –if you understand me- , you can simply b4 playing any game make sure that it will not make the other wins.

But estimation you can never define how the opponent will win, so it need you  to be more intelligent

 

One nice thing is that you don’t have only  one opponent , you have 3 opponents you have to deal with, and your turn always change, sometimes you are the first to play sometimes the last.

 

And the most interesting thing that the 4 groups will play against each other in one game .

 

So I think it is a nice idea

 

About the penalties , I can’t believe it is an AI project :S

[Dooma®]

First, I add my voice to the Penalties haters. It is time wasting to
implement graphics and networks and leave the AI agent
Sherif, I totally agree with you on the estimation project.
For the chatting bot, i think it is nice too, although i have no clue
how it will be done :o)
but for judging, I was thinking also of Turing test. I don't think
there is other ways of judging the agents.

[Douma©]

The football game project is most likely an EA Sports game that "Challenge Everything"!!!!

There is a wide variety of projects that could help us better to show our AI skills. One of them is the Estimation choice which I am supporting for many reasons.

• It needs really a smart AI agent.
• The evaluation will be obviously clear.        
• There will be no space for augmenting the scope of the project :D ( I mean that a tremendous estimation game will not need a complicated graphics engine ).

I am suggesting also a Chess game( i believe it's the best AI game but the hardest one ) , Backgammon, Checkers...etc.

I mean look at sports, there are many sports where there is no direct competition between competitiors. Actually there is no contact between them at all. Everyone just does his/her part and the jury gives their mark. 

I agree with you, but juries gives there marks based on some standards and rules, if we can set such rules that will be sufficient for evaluating the best conversation, so why not!!!
But probably doctors choose competition to simplify the evaluation process.

--
Adham Tammam

[Mahmoud Shalaby]

Salamo 3alikom,

Sorry bas I don't like to have a looooong e-mail we fel akher my comment so that's why I deleted what was written first. I assume that all of you know what is this anyhow :). I just wanted also to add something concerning the estimation game, the agent will also must decide the number of rounds to collect (the bid, the call, el mazad, ... whatever name). However, this game will need a FIFTH agent that will control the game, it should check that the players do not cheat, make up the scores, accept the bids, ... etc. However, I really do appreciate this idea much more on the football thing ...

Salamo 3alikom,

--
Thanks and Best Regards,
Mahmoud M. Shalaby
IT and Publications Head, STP 2007
"Developing our Community"
Visit: http://www.stp-egypt.com/

Engineering is applying math and science to create useful things.
Engineering is looking forward, seeing what the world needs, and
making it happen. Engineering is making people's lives easier,
safer and better through technology. What kind of future will you create ??

[Ahmed Mostafa]

assalamo 3alaykom,

 

I don't know what the "Estimation" game is, Mahammad Mohsen will tell me later.
Anyway, I prefer connect-4

 

Salam
 

Faculty of Engineering
Cairo University

[Ali Helmy]

On 10/10/06, Douma© <adh...@gmail.com> wrote:

On 10/10/06, Dooma® <adham.a...@gmail.com > wrote:

I am suggesting also a Chess game( i believe it's the best AI game but the hardest one ) , Backgammon, Checkers...etc.

--
Adham Tammam

Well, i agree that chess is very nice too, but the problem is, last year they had chess as their problem, so probably nevein doesnt want to risk Copy+Paste issues