On 7 févr. 2013, at 21:01, Olivier Teytaud wrote:
> (the constant term in the linear slope in log-log term depends on the
> number of parameters,
> this might be relevant for Brian's remark on the number of parameters).
I am not sure it does. In fact I would guess it does not. If the opt
I have a rationale for choosing "3 or 4" as the optimum. Not theoretically
sound or anything, but this is how I am actually looking at.
The number of parameters in a Clop model with N input variables is (N + 1) *
(N + 2) / 2. Here is a table:
1 --> 3
2 --> 6
3 --> 10
4 --> 15
5 --> 21
Thanks Brian for this comment:
> 1) It is very important to limit the number of parameters concurrently
> optimized. E.g., to 4 or fewer. The optimum might be 3.
>
I do something similar (with other optimization algorithms); in large
dimension, I optimize only a few parameters,
and then I slowly
Thanks to you both for feedback;
The CLOP paper has references to papers with these bounds (Chen-1988 for
> the proof that it is not possible to do better. IIRC, some variations of
> stochastic gradient have been proved to match the bound, so the bound is
> tight).
>
> Yes, Chen's bound is the mai
If I understand your notation correctly, CLOP can get C = 1/3 for smooth
functions.
In theory, regression with a polynomial of degree d gets C = 0.5 * d / (d + 1)
if the function is as smooth as the polynomial with bounded (d+1)th derivative.
There is an experiment in the CLOP paper that confir
I use Clop for optimizing a lot of Pebbles parameters. But I can't qualify
performance in the way you describe, because I actually don't know what the
optimum is.
I can relate some observations from using it:
0) Many parameters are easier to optimize using other systems. E.g., Remi
uses mi
Ofcourse a longer thinking time can yield a worse move.
That's like a weaker player playing a better move than a stronger
player would have. It only means that the position is too complicated
for both of them, and the weaker player got lucky.
___
Computer
Hi;
as many of you I guess, I often try to optimize parameters of strategies.
In games, if I optimize against a fixed opponent, this is a noisy
optimization problem: I want the parameters with the best success rate
against the chosen opponent,
which can be evaluated thanks to repeated (time-consu
Dear All,
In the previous mail, I wrote that the prizes are:
1st, 30, yen; 2nd, 20, yen; 3rd 5, yen.
The position of commas are according to Japanese counting way of numbers.
It is pointed out by Okazaki san that in the Western country, they use:
1st, 300,000 yen; 2nd, 200,000 yen; 3
Dear All,
Finally, the web page of the 6th UEC Cup Computer Go Tournament is open!
Please check at: http://jsb.cs.uec.ac.jp/~igo/eng/
Registration is also open now.
We have a big CHANGE in the UEC Cup this year.
Below I excerpt some information on change of the 6th UEC Cup, but do not
trust the
10 matches
Mail list logo