Jump to content

Auto-WEKA

From Wikipedia, the free encyclopedia

Auto-WEKA is an automated machine learning system based on WEKA by Chris Thornton, Frank Hutter, Holger H. Hoos and Kevin Leyton-Brown.[1] An extended version was published as Auto-WEKA 2.0.[2]

Description

[edit]

It received the test-of-time award of the SIGKDD conference in 2023.[3]

Auto-WEKA introduced the Combined Algorithm Selection and Hyperparameter optimization (CASH) problem, a formalism that was picked up by later AutoML systems such as Auto-sklearn. It extends both the Algorithm selection problem and the Hyperparameter optimization problem, by searching for the best algorithm and also its hyperparameters for a given dataset.[citation needed]

References

[edit]
  1. ^ Thornton, Chris; Hutter, Frank; Hoos, Holger H.; Leyton-Brown, Kevin (August 11, 2013). "Auto-WEKA: combined selection and hyperparameter optimization of classification algorithms". Association for Computing Machinery. pp. 847–855. doi:10.1145/2487575.2487629 – via ACM Digital Library.
  2. ^ Kotthoff, Lars; Thornton, Chris; Hoos, Holger H.; Hutter, Frank; Leyton-Brown, Kevin (August 12, 2017). "Auto-WEKA 2.0: Automatic model selection and hyperparameter optimization in WEKA". Journal of Machine Learning Research. 18 (25): 1–5 – via jmlr.org.
  3. ^ "KDD 2023 - Awards". kdd.org.