Artificial Intelligence Depot
Visiting guest. Why not sign in?
News, knowledge and discussion for the AI enthusiast.
FEATURES COMMUNITY KNOWLEDGE SEARCH  
Data Mining: Practical Machine Learning Tools and Techniques
Data mining techniques are used to power intelligent software. The book shows how machine learning can be used to find predictive patterns in data comprehensible to managers and developers alike.
More information at Amazon US UK

Reply to Message

Not registered yet?

The AI Depot has a focused community of friendly users. Rather than let anyone abuse the site at the brink of promiscuity, we prefer to let only those with an active interest participate... this simply requires registering.

Why not sign up!

Joining the site's community is completely free. You can then post messages freely, and customise your personal profile at will. Specific privileges will also be granted to you, like being able to access printer-friendly articles without restrictions. So, why not register?

Username:
Password:
Subject:
Email me when someone replies.
Body:

Parent Message

Googly Contest!

For those of you that have not yet heard the news, Google is holding their first annual contest in order to get lots of amazing code from the community ;) However, you can get money out of it in exchange! See the Google Programming Contest Page for more information.

I've already taken a look at the code and some of the data files, at it seems a very good platform for NLP experimentation. They provide all the basic processing tools you need, but its up to you to build upon that. It's definitely worth looking into, even if you don't join the contest -- though you should watch out for the license terms!

935 posts.
Friday 08 February, 12:39
Reply
Contest rules

They don't seem completely unreasonable though. As I understand it, you retain the rights to the code/algorithms etc, you just give Google the right to use it in whatever way they see fit. This does not prevent you from selling your technology to other parties.

Of course a risk is that you will find Google as your competitor, but they are probably not interested in selling software-components/technologies.

Jeroen

28 posts.
Friday 08 February, 12:58
Reply
Legal Issues

It's not so much that, but I'm not sure if you're actually allowed to use the source when not entering the contest... it says for 'academic' and 'research' purposes; so I suppose it's pretty open!

935 posts.
Saturday 09 February, 08:09
Reply
Heh. Nice try Google!

Nice try Google! Letting others do their own work! ;)
I'd say, $10.000 is not much for such a technology. Bet you could sell it better to Carnegie Mellon (Lycos) or Microsoft.

But for the compression part, I already have a damn cool idea... which has yet to be tested. And, AFAIK it hasn't been used yet (which is the reason for my doubts... but it has to work! theoretically at least...)

The license part needs clarification, though.

-Johannes

13 posts.
Sunday 10 February, 09:34
Reply
Ideal Testbed

I think it's good. Yeah, Google benefits; they get a lot of PR and stand to get a fair few entries. That said, I don't expect many to be that great... as you said, people with brains will not give away their code that easily ;)

But from the other point of view, I admire them for giving out that much code, and a lot of very useful test data. That's damn cool, and many AI hobbyist will have a great time experimenting.

On the other hand, if you submit and win, you get instant world fame a bit of money, and some technology that everyone else will want. Not to mention people wanting to give you very well payed jobs!

May be worth it... good luck with your idea -- any clues what it does? ;)

935 posts.
Sunday 10 February, 13:15
Reply
Clue ;)

It's a compression algorithm using the only mathematical method to represent an infinite amount of data with a fitite one. Note that the data to compress can only be infinite in theory.
Happy riddling ;)

As for the code, I haven't looked at it. May well be interesting.
Too bad I can't take part in the contest (age). Wouldn't be much difference, though, since I won't make it far :)

13 posts.
Monday 11 February, 17:11
Reply
Fractals!

Of course... well if you can apply that to web-page compression then you're a better person than I am. My knowledge of fractals is limited to drawing them and zooming in on them!

If the age thing is the only thing bothering you, I'm sure you can get someone older to submit! Don't let that stop your experimentation ;)

935 posts.
Monday 11 February, 17:31
Reply
Nope!

Actually I'm not talking about Fractals, but of Functions. By finding an explicit form for a series of numbers, you can pack them (but i'm sure you already knew that). This is actually done in RLE but only with constant series, afaik. By using non-polynomial functions (i.e. integer) this must be quite well working because you only have to approximate the function and then apply the non-polynomial function. (I hope you get what I mean)

I know that the compressing parts may take hours, but it may well be worth the effort. Bet you'll need damn much simplification algorithms and graph theory for implementation.

Maybe (or: most probably) this is just another crazy thought of mine. But hey, I just can't figure out why this can't be realized / is not effective enough.

13 posts.
Wednesday 13 February, 11:35
Reply

Back to the Artificial Intelligence Depot.