Matches in SemOpenAlex for { <https://semopenalex.org/work/W3202444716> ?p ?o ?g. }
Showing items 1 to 83 of
83
with 100 items per page.
- W3202444716 startingPage "16" @default.
- W3202444716 abstract "Greedy decision tree learning heuristics are mainstays of machine learning practice, but theoretical justification for their empirical success remains elusive. In fact, it has long been known that there are simple target functions for which they fail badly (Kearns and Mansour, STOC 1996).Recent work of Brutzkus, Daniely, and Malach (COLT 2020) considered the smoothed analysis model as a possible avenue towards resolving this disconnect. Within the smoothed setting and for targets f that are k-juntas, they showed that these heuristics successfully learn f with depth-k decision tree hypotheses. They conjectured that the same guarantee holds more generally for targets that are depth-k decision trees. We provide a counterexample to this conjecture: we construct targets that are depth-k decision trees and show that even in the smoothed setting, these heuristics build trees of depth 2^{Ω(k)} before achieving high accuracy. We also show that the guarantees of Brutzkus et al. cannot extend to the agnostic setting: there are targets that are very close to k-juntas, for which these heuristics build trees of depth 2^{Ω(k)} before achieving high accuracy." @default.
- W3202444716 created "2021-10-11" @default.
- W3202444716 creator A5011657876 @default.
- W3202444716 creator A5023118987 @default.
- W3202444716 creator A5029451171 @default.
- W3202444716 creator A5059516326 @default.
- W3202444716 date "2021-01-01" @default.
- W3202444716 modified "2023-10-07" @default.
- W3202444716 title "Decision Tree Heuristics Can Fail, Even in the Smoothed Setting" @default.
- W3202444716 doi "https://doi.org/10.4230/lipics.approx/random.2021.45" @default.
- W3202444716 hasPublicationYear "2021" @default.
- W3202444716 type Work @default.
- W3202444716 sameAs 3202444716 @default.
- W3202444716 citedByCount "0" @default.
- W3202444716 crossrefType "proceedings-article" @default.
- W3202444716 hasAuthorship W3202444716A5011657876 @default.
- W3202444716 hasAuthorship W3202444716A5023118987 @default.
- W3202444716 hasAuthorship W3202444716A5029451171 @default.
- W3202444716 hasAuthorship W3202444716A5059516326 @default.
- W3202444716 hasConcept C111472728 @default.
- W3202444716 hasConcept C113174947 @default.
- W3202444716 hasConcept C11413529 @default.
- W3202444716 hasConcept C114614502 @default.
- W3202444716 hasConcept C119857082 @default.
- W3202444716 hasConcept C126255220 @default.
- W3202444716 hasConcept C127705205 @default.
- W3202444716 hasConcept C138885662 @default.
- W3202444716 hasConcept C154945302 @default.
- W3202444716 hasConcept C162838799 @default.
- W3202444716 hasConcept C199360897 @default.
- W3202444716 hasConcept C2780586882 @default.
- W3202444716 hasConcept C2780801425 @default.
- W3202444716 hasConcept C2780990831 @default.
- W3202444716 hasConcept C33923547 @default.
- W3202444716 hasConcept C41008148 @default.
- W3202444716 hasConcept C80444323 @default.
- W3202444716 hasConcept C84525736 @default.
- W3202444716 hasConceptScore W3202444716C111472728 @default.
- W3202444716 hasConceptScore W3202444716C113174947 @default.
- W3202444716 hasConceptScore W3202444716C11413529 @default.
- W3202444716 hasConceptScore W3202444716C114614502 @default.
- W3202444716 hasConceptScore W3202444716C119857082 @default.
- W3202444716 hasConceptScore W3202444716C126255220 @default.
- W3202444716 hasConceptScore W3202444716C127705205 @default.
- W3202444716 hasConceptScore W3202444716C138885662 @default.
- W3202444716 hasConceptScore W3202444716C154945302 @default.
- W3202444716 hasConceptScore W3202444716C162838799 @default.
- W3202444716 hasConceptScore W3202444716C199360897 @default.
- W3202444716 hasConceptScore W3202444716C2780586882 @default.
- W3202444716 hasConceptScore W3202444716C2780801425 @default.
- W3202444716 hasConceptScore W3202444716C2780990831 @default.
- W3202444716 hasConceptScore W3202444716C33923547 @default.
- W3202444716 hasConceptScore W3202444716C41008148 @default.
- W3202444716 hasConceptScore W3202444716C80444323 @default.
- W3202444716 hasConceptScore W3202444716C84525736 @default.
- W3202444716 hasLocation W32024447161 @default.
- W3202444716 hasOpenAccess W3202444716 @default.
- W3202444716 hasPrimaryLocation W32024447161 @default.
- W3202444716 hasRelatedWork W1492170170 @default.
- W3202444716 hasRelatedWork W1967684610 @default.
- W3202444716 hasRelatedWork W2040973615 @default.
- W3202444716 hasRelatedWork W2050660892 @default.
- W3202444716 hasRelatedWork W2130186484 @default.
- W3202444716 hasRelatedWork W2147306258 @default.
- W3202444716 hasRelatedWork W2150181672 @default.
- W3202444716 hasRelatedWork W2155831576 @default.
- W3202444716 hasRelatedWork W2224634587 @default.
- W3202444716 hasRelatedWork W2736121450 @default.
- W3202444716 hasRelatedWork W2963865651 @default.
- W3202444716 hasRelatedWork W2964247925 @default.
- W3202444716 hasRelatedWork W3023883850 @default.
- W3202444716 hasRelatedWork W3094727079 @default.
- W3202444716 hasRelatedWork W3107849711 @default.
- W3202444716 hasRelatedWork W3124955340 @default.
- W3202444716 hasRelatedWork W3180727286 @default.
- W3202444716 hasRelatedWork W3206488636 @default.
- W3202444716 hasRelatedWork W3208785465 @default.
- W3202444716 hasRelatedWork W3212747081 @default.
- W3202444716 isParatext "false" @default.
- W3202444716 isRetracted "false" @default.
- W3202444716 magId "3202444716" @default.
- W3202444716 workType "article" @default.