4 ms·

Shrinking bull’s-eye algorithm speeds up complex modeling from days to hours

33 points by chmaynard 11y ago

dzdt 11y ago
Sounds like simulated annealing described in a simplified-for-the-journalist way. Anyone know what the new thing really is?
- spooningtamarin 11y ago
  How do you determine which is the best next point where to evaluate the model? Simulated annealing is blind, what you do with simulated annealing you just move locally around the current point. If you had a partial picture of the landscape of the function, how would you locate the most interesting unvisited point in the landscape? I do have to admit that what I'm seeing here is only the application of Bayesian sampling algorithms, there are startups doing similar things for A/B testing and similar.
- stkni 11y ago
  Original paper was linked from article and is here [1]. Although I'm vaguely familiar with MCMC and Metropolis, it's not nearly enough to pass comment on this advance. [1] http://arxiv.org/pdf/1402.1694v4.pdf http://arxiv.org/pdf/1402.1694v4.pdf
  - misterthirsty 11y ago
    The significant change in Bayesian methods is moving away from Metropolis Hastings to an approach based on Hamiltonian dynamics, implemented widely in the STAN package. It would be interesting to see if this is what is being used.
- brudgers 11y ago
  My understanding of simulated annealing is that the "fudge factor" is chosen and reduced more or less "naively". By which I mean it's initial value and reduction across iterations is hard wired into the execution path before it runs and...the parameters for iteration \i\ are determined before iteration \1\. My understanding of the article is that the new algorithm determines the parameters of iteration \i\ statistically from the results of iterations \1\ through \i-1\.
- kwhitefoot 11y ago
  From a cursory reading of the article it sounds like they optimize for the best annealling schedule which of course is what every manufacturer of steel has been doing with varying degrees of success for at least a thousand years.
alexholehouse 11y ago
From a cursory reading of the article this looks like it's solving an extremely similar problem to Gaussian Process Bayesian Optimization (GBPO [1]) in a fairly similar way. I wonder what a head-to-head comparison between these two would look like. [1] J. Gardner, M. Kusner, Z. Xu, K. Weinberger, and J. Cunningham, in Proceedings of the 31st International Conference on Machine Learning (ICML-14) (JMLR Workshop and Conference Proceedings, 2014), p. 937
- spooningtamarin 11y ago
  Yep, MOE [1] is an example of an open source GPBO. [1] https://github.com/sigopt/MOE https://github.com/sigopt/MOE
arbitrage314 11y ago
I don't understand this paper yet, but Marzouk is an awesome fella--a great explainer, a brilliant researcher, etc. I almost took a job with him a few years ago. If I get through this paper quick enough, I'll post back any significant thoughts I may have (if any).

33 points by chmaynard 11y ago

dzdt 11y ago

Sounds like simulated annealing described in a simplified-for-the-journalist way. Anyone know what the new thing really is?

spooningtamarin 11y ago
How do you determine which is the best next point where to evaluate the model? Simulated annealing is blind, what you do with simulated annealing you just move locally around the current point. If you had a partial picture of the landscape of the function, how would you locate the most interesting unvisited point in the landscape? I do have to admit that what I'm seeing here is only the application of Bayesian sampling algorithms, there are startups doing similar things for A/B testing and similar.
stkni 11y ago
Original paper was linked from article and is here [1]. Although I'm vaguely familiar with MCMC and Metropolis, it's not nearly enough to pass comment on this advance. [1] http://arxiv.org/pdf/1402.1694v4.pdf http://arxiv.org/pdf/1402.1694v4.pdf
- misterthirsty 11y ago
  The significant change in Bayesian methods is moving away from Metropolis Hastings to an approach based on Hamiltonian dynamics, implemented widely in the STAN package. It would be interesting to see if this is what is being used.
brudgers 11y ago
My understanding of simulated annealing is that the "fudge factor" is chosen and reduced more or less "naively". By which I mean it's initial value and reduction across iterations is hard wired into the execution path before it runs and...the parameters for iteration \i\ are determined before iteration \1\. My understanding of the article is that the new algorithm determines the parameters of iteration \i\ statistically from the results of iterations \1\ through \i-1\.
kwhitefoot 11y ago
From a cursory reading of the article it sounds like they optimize for the best annealling schedule which of course is what every manufacturer of steel has been doing with varying degrees of success for at least a thousand years.

alexholehouse 11y ago

From a cursory reading of the article this looks like it's solving an extremely similar problem to Gaussian Process Bayesian Optimization (GBPO [1]) in a fairly similar way. I wonder what a head-to-head comparison between these two would look like. [1] J. Gardner, M. Kusner, Z. Xu, K. Weinberger, and J. Cunningham, in Proceedings of the 31st International Conference on Machine Learning (ICML-14) (JMLR Workshop and Conference Proceedings, 2014), p. 937

spooningtamarin 11y ago
Yep, MOE [1] is an example of an open source GPBO. [1] https://github.com/sigopt/MOE https://github.com/sigopt/MOE

arbitrage314 11y ago

I don't understand this paper yet, but Marzouk is an awesome fella--a great explainer, a brilliant researcher, etc. I almost took a job with him a few years ago. If I get through this paper quick enough, I'll post back any significant thoughts I may have (if any).