We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

cs.LG

Change to browse by:

References & Citations

DBLP - CS Bibliography

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Machine Learning

Title: ExperienceThinking: Constrained Hyperparameter Optimization based on Knowledge and Pruning

Abstract: Machine learning algorithms are very sensitive to the hyperparameters, and their evaluations are generally expensive. Users desperately need intelligent methods to quickly optimize hyperparameter settings according to known evaluation information, and thus reduce computational cost and promote optimization efficiency. Motivated by this, we propose ExperienceThinking algorithm to quickly find the best possible hyperparameter configuration of machine learning algorithms within a few configuration evaluations. ExperienceThinking design two novel methods, which intelligently infer optimal configurations from two aspects: search space pruning and knowledge utilization respectively. Two methods complement each other and solve the constrained hyperparameter optimization problems effectively. To demonstrate the benefit of ExperienceThinking, we compare it with 3 classical hyperparameter optimization algorithms with a small number of configuration evaluations. The experimental results present that our proposed algorithm provides superior results and achieve better performance.
Subjects: Machine Learning (cs.LG); Machine Learning (stat.ML)
Cite as: arXiv:1912.00602 [cs.LG]
  (or arXiv:1912.00602v2 [cs.LG] for this version)

Submission history

From: Chunnnan Wang [view email]
[v1] Mon, 2 Dec 2019 07:21:05 GMT (368kb,D)
[v2] Mon, 4 May 2020 05:48:17 GMT (870kb,D)

Link back to: arXiv, form interface, contact.