Matches in SemOpenAlex for { <https://semopenalex.org/work/W2895932056> ?p ?o ?g. }
- W2895932056 abstract "Popular machine learning estimators involve regularization parameters that can be challenging to tune, and standard strategies rely on grid search for this task. In this paper, we revisit the techniques of approximating the regularization path up to predefined tolerance $epsilon$ in a unified framework and show that its complexity is $O(1/sqrt[d]{epsilon})$ for uniformly convex loss of order $d geq 2$ and $O(1/sqrt{epsilon})$ for Generalized Self-Concordant functions. This framework encompasses least-squares but also logistic regression, a case that as far as we know was not handled as precisely in previous works. We leverage our technique to provide refined bounds on the validation error as well as a practical algorithm for hyperparameter tuning. The latter has global convergence guarantee when targeting a prescribed accuracy on the validation set. Last but not least, our approach helps relieving the practitioner from the (often neglected) task of selecting a stopping criterion when optimizing over the training set: our method automatically calibrates this criterion based on the targeted accuracy on the validation set." @default.
- W2895932056 created "2018-10-26" @default.
- W2895932056 creator A5021079415 @default.
- W2895932056 creator A5033768552 @default.
- W2895932056 creator A5064454585 @default.
- W2895932056 creator A5081482638 @default.
- W2895932056 creator A5090184417 @default.
- W2895932056 date "2018-10-12" @default.
- W2895932056 modified "2023-10-01" @default.
- W2895932056 title "Safe Grid Search with Optimal Complexity" @default.
- W2895932056 cites W1528905581 @default.
- W2895932056 cites W1543911290 @default.
- W2895932056 cites W1552009253 @default.
- W2895932056 cites W1556545893 @default.
- W2895932056 cites W2022004743 @default.
- W2895932056 cites W2047028564 @default.
- W2895932056 cites W205960364 @default.
- W2895932056 cites W2062102668 @default.
- W2895932056 cites W2063978378 @default.
- W2895932056 cites W2097360283 @default.
- W2895932056 cites W2101095383 @default.
- W2895932056 cites W2101234009 @default.
- W2895932056 cites W2109706083 @default.
- W2895932056 cites W2113651538 @default.
- W2895932056 cites W2122825543 @default.
- W2895932056 cites W2133958955 @default.
- W2895932056 cites W2135046866 @default.
- W2895932056 cites W2146258363 @default.
- W2895932056 cites W2173251738 @default.
- W2895932056 cites W2269622027 @default.
- W2895932056 cites W2312732607 @default.
- W2895932056 cites W2603538444 @default.
- W2895932056 cites W2625093932 @default.
- W2895932056 cites W2804529069 @default.
- W2895932056 cites W2906263529 @default.
- W2895932056 cites W2962685267 @default.
- W2895932056 cites W2962826047 @default.
- W2895932056 cites W2962939576 @default.
- W2895932056 cites W2963319429 @default.
- W2895932056 cites W2963518220 @default.
- W2895932056 cites W2963576992 @default.
- W2895932056 cites W2963815651 @default.
- W2895932056 cites W3104887532 @default.
- W2895932056 cites W918148966 @default.
- W2895932056 hasPublicationYear "2018" @default.
- W2895932056 type Work @default.
- W2895932056 sameAs 2895932056 @default.
- W2895932056 citedByCount "0" @default.
- W2895932056 crossrefType "posted-content" @default.
- W2895932056 hasAuthorship W2895932056A5021079415 @default.
- W2895932056 hasAuthorship W2895932056A5033768552 @default.
- W2895932056 hasAuthorship W2895932056A5064454585 @default.
- W2895932056 hasAuthorship W2895932056A5081482638 @default.
- W2895932056 hasAuthorship W2895932056A5090184417 @default.
- W2895932056 hasConcept C10485038 @default.
- W2895932056 hasConcept C105795698 @default.
- W2895932056 hasConcept C112680207 @default.
- W2895932056 hasConcept C11413529 @default.
- W2895932056 hasConcept C12267149 @default.
- W2895932056 hasConcept C126255220 @default.
- W2895932056 hasConcept C153083717 @default.
- W2895932056 hasConcept C154945302 @default.
- W2895932056 hasConcept C185429906 @default.
- W2895932056 hasConcept C187691185 @default.
- W2895932056 hasConcept C2524010 @default.
- W2895932056 hasConcept C2776135515 @default.
- W2895932056 hasConcept C33923547 @default.
- W2895932056 hasConcept C41008148 @default.
- W2895932056 hasConcept C8642999 @default.
- W2895932056 hasConceptScore W2895932056C10485038 @default.
- W2895932056 hasConceptScore W2895932056C105795698 @default.
- W2895932056 hasConceptScore W2895932056C112680207 @default.
- W2895932056 hasConceptScore W2895932056C11413529 @default.
- W2895932056 hasConceptScore W2895932056C12267149 @default.
- W2895932056 hasConceptScore W2895932056C126255220 @default.
- W2895932056 hasConceptScore W2895932056C153083717 @default.
- W2895932056 hasConceptScore W2895932056C154945302 @default.
- W2895932056 hasConceptScore W2895932056C185429906 @default.
- W2895932056 hasConceptScore W2895932056C187691185 @default.
- W2895932056 hasConceptScore W2895932056C2524010 @default.
- W2895932056 hasConceptScore W2895932056C2776135515 @default.
- W2895932056 hasConceptScore W2895932056C33923547 @default.
- W2895932056 hasConceptScore W2895932056C41008148 @default.
- W2895932056 hasConceptScore W2895932056C8642999 @default.
- W2895932056 hasLocation W28959320561 @default.
- W2895932056 hasOpenAccess W2895932056 @default.
- W2895932056 hasPrimaryLocation W28959320561 @default.
- W2895932056 hasRelatedWork W2100268245 @default.
- W2895932056 hasRelatedWork W2119624713 @default.
- W2895932056 hasRelatedWork W2186459748 @default.
- W2895932056 hasRelatedWork W2773043772 @default.
- W2895932056 hasRelatedWork W2784329877 @default.
- W2895932056 hasRelatedWork W2794754002 @default.
- W2895932056 hasRelatedWork W2808560264 @default.
- W2895932056 hasRelatedWork W2907303449 @default.
- W2895932056 hasRelatedWork W2942460262 @default.
- W2895932056 hasRelatedWork W2943296448 @default.
- W2895932056 hasRelatedWork W2946765615 @default.
- W2895932056 hasRelatedWork W2952131993 @default.
- W2895932056 hasRelatedWork W2963172602 @default.