CLOP

Home * Automated Tuning * CLOP

CLOP for Noisy Black-Box Parameter Optimization, a program used for parameter tuning, created by Rémi Coulom. The CLOP principle, which stands for Confident Local OPtimization [1], is an approach to local regression that overcomes problems with very noisy outputs and nonnegative Hessians [2] in a simple and efficient way. CLOP is suited to tune several parameters simultaneously, and requires polynomial time for each additional parameter to verify the tuned values, that is playing a lot of games [3].

See also

Publications

Forum Posts

2011 …

2012

2013

2014

2015 …

2020 …

References

  1. CLOP: Confident Local Optimization for Noisy Black-Box Parameter Tuning
  2. Second partial derivative test from Wikipedia
  3. Re: How Do You Automatically Tune Your Evaluation Tables by Jon Dart, CCC, January 31, 2014
  4. The Inventions - Les Paul

Up one Level