MATLAB Central Spring Contest
in [Matlab]
|
From: Jin on 7 May 2008 02:46 why bombing the queue? I sugguest adding some submission limits, such as 1 entry/per min/per ip.
From: Jin on 7 May 2008 02:49 why bombing? when is the end?
From: Markus Buehren on 7 May 2008 04:23 Alans bombing as well as his analysis both speak for adding a "captcha" to the submission site (which I think is even better than an IP limit). This would prevent the over-over-over-specification of the solver to the test suite which occurs due to the many parameter variations that are submitted. I have found some algorithmic improvement yesterday, which dramatically boosted the score on the test suite we are having at home. However, as the leading code is so over-specified to the actual test suite, my change even lead to a worse score on it. I think the parameter tweaking by numerous automated uploads takes away the room for algorithmic improvements. The most obvious sign for tweaking is the introduction of randomness into the algorithm. Not only that parameters of the code are varied, we even don't know which parameters that are! In my opinion the use of random generators (rand, randn, randperm etc.) should be forbidden. Any other thoughts? Markus
From: srach on 7 May 2008 04:47 > I have found some algorithmic improvement yesterday, which > dramatically boosted the score on the test suite we are > having at home. However, as the leading code is so > over-specified to the actual test suite, my change even lead > to a worse score on it. I think the parameter tweaking by > numerous automated uploads takes away the room for > algorithmic improvements. But sometimes such differences also come with different matlab versions or different machines. Earlier in the contest, I found a small change that led to an improvement of about 10 sec on my computer. I checked that on a different computer (with a more recent matlab version) and there this change led to a worsening of 10 sec. (Same was true at the end of the last contest, where the final tweak that I saved for the last seconds of the contest actually made the score worse, although it led to improvements on my system.) I second your thoughts about randomness. Regards, srach
From: srach on 7 May 2008 04:47
> I have found some algorithmic improvement yesterday, which > dramatically boosted the score on the test suite we are > having at home. However, as the leading code is so > over-specified to the actual test suite, my change even lead > to a worse score on it. I think the parameter tweaking by > numerous automated uploads takes away the room for > algorithmic improvements. But sometimes such differences also come with different matlab versions or different machines. Earlier in the contest, I found a small change that led to an improvement of about 10 sec on my computer. I checked that on a different computer (with a more recent matlab version) and there this change led to a worsening of 10 sec. (Same was true at the end of the last contest, where the final tweak that I saved for the last seconds of the contest actually made the score worse, although it led to improvements on my system.) I second your thoughts about randomness. Regards, srach |