Matches in SemOpenAlex for { <https://semopenalex.org/work/W2989742941> ?p ?o ?g. }
- W2989742941 endingPage "98" @default.
- W2989742941 startingPage "86" @default.
- W2989742941 abstract "We study the complexity of learning and approximation of self-bounding functions over the uniform distribution on the Boolean hypercube {0,1}n. Informally, a function f:{0,1}n→R is self-bounding if for every x∈{0,1}n, f(x) upper bounds the sum of all the n marginal decreases in the value of the function at x. Self-bounding functions include such well-known classes of functions as submodular and fractionally-subadditive (XOS) functions. They were introduced by Boucheron et al. (2010) in the context of concentration of measure inequalities. Our main result is a nearly tight ℓ1-approximation of self-bounding functions by low-degree juntas. Specifically, all self-bounding functions can be ϵ-approximated in ℓ1 by a polynomial of degree O˜(1/ϵ) over 2O˜(1/ϵ) variables. We show that both the degree and junta-size are optimal up to logarithmic terms. Previous techniques considered stronger ℓ2 approximation and proved nearly tight bounds of Θ(1/ϵ2) on the degree and 2Θ(1/ϵ2) on the number of variables. Our bounds rely on the analysis of noise stability of self-bounding functions together with a stronger connection between noise stability and ℓ1 approximation by low-degree polynomials. This technique can also be used to get tighter bounds on ℓ1 approximation by low-degree polynomials and a faster learning algorithm for halfspaces. These results lead to improved and in several cases almost tight bounds for PAC and agnostic learning of self-bounding functions relative to the uniform distribution. In particular, assuming hardness of learning juntas, we show that PAC and agnostic learning of self-bounding functions have complexity of nΘ˜(1/ϵ)." @default.
- W2989742941 created "2019-12-05" @default.
- W2989742941 creator A5023410548 @default.
- W2989742941 creator A5054675941 @default.
- W2989742941 creator A5068103506 @default.
- W2989742941 date "2020-02-01" @default.
- W2989742941 modified "2023-09-25" @default.
- W2989742941 title "Tight bounds on ℓ1 approximation and learning of self-bounding functions" @default.
- W2989742941 cites W1489827011 @default.
- W2989742941 cites W1642458096 @default.
- W2989742941 cites W1779941205 @default.
- W2989742941 cites W1897230935 @default.
- W2989742941 cites W1978769275 @default.
- W2989742941 cites W1982381767 @default.
- W2989742941 cites W1986296546 @default.
- W2989742941 cites W1995275762 @default.
- W2989742941 cites W2019121884 @default.
- W2989742941 cites W2019363670 @default.
- W2989742941 cites W2023683873 @default.
- W2989742941 cites W2040884411 @default.
- W2989742941 cites W2042468189 @default.
- W2989742941 cites W2063003848 @default.
- W2989742941 cites W2064680241 @default.
- W2989742941 cites W2084544490 @default.
- W2989742941 cites W2103749128 @default.
- W2989742941 cites W2106458073 @default.
- W2989742941 cites W2121372565 @default.
- W2989742941 cites W2129192653 @default.
- W2989742941 cites W2147306258 @default.
- W2989742941 cites W2148858749 @default.
- W2989742941 cites W2157316274 @default.
- W2989742941 cites W2160354932 @default.
- W2989742941 cites W2163032969 @default.
- W2989742941 cites W2460132610 @default.
- W2989742941 cites W2953018665 @default.
- W2989742941 cites W2962711478 @default.
- W2989742941 cites W2963042078 @default.
- W2989742941 cites W2963394384 @default.
- W2989742941 cites W2963650730 @default.
- W2989742941 cites W2963798205 @default.
- W2989742941 cites W2963974416 @default.
- W2989742941 cites W2964184416 @default.
- W2989742941 cites W48624319 @default.
- W2989742941 doi "https://doi.org/10.1016/j.tcs.2019.11.013" @default.
- W2989742941 hasPublicationYear "2020" @default.
- W2989742941 type Work @default.
- W2989742941 sameAs 2989742941 @default.
- W2989742941 citedByCount "2" @default.
- W2989742941 countsByYear W29897429412021 @default.
- W2989742941 crossrefType "journal-article" @default.
- W2989742941 hasAuthorship W2989742941A5023410548 @default.
- W2989742941 hasAuthorship W2989742941A5054675941 @default.
- W2989742941 hasAuthorship W2989742941A5068103506 @default.
- W2989742941 hasConcept C114614502 @default.
- W2989742941 hasConcept C118615104 @default.
- W2989742941 hasConcept C121332964 @default.
- W2989742941 hasConcept C134306372 @default.
- W2989742941 hasConcept C14036430 @default.
- W2989742941 hasConcept C148764684 @default.
- W2989742941 hasConcept C154945302 @default.
- W2989742941 hasConcept C178621042 @default.
- W2989742941 hasConcept C187455244 @default.
- W2989742941 hasConcept C19220575 @default.
- W2989742941 hasConcept C24890656 @default.
- W2989742941 hasConcept C2775997480 @default.
- W2989742941 hasConcept C33923547 @default.
- W2989742941 hasConcept C41008148 @default.
- W2989742941 hasConcept C50820777 @default.
- W2989742941 hasConcept C63584917 @default.
- W2989742941 hasConcept C77553402 @default.
- W2989742941 hasConcept C78458016 @default.
- W2989742941 hasConcept C86803240 @default.
- W2989742941 hasConceptScore W2989742941C114614502 @default.
- W2989742941 hasConceptScore W2989742941C118615104 @default.
- W2989742941 hasConceptScore W2989742941C121332964 @default.
- W2989742941 hasConceptScore W2989742941C134306372 @default.
- W2989742941 hasConceptScore W2989742941C14036430 @default.
- W2989742941 hasConceptScore W2989742941C148764684 @default.
- W2989742941 hasConceptScore W2989742941C154945302 @default.
- W2989742941 hasConceptScore W2989742941C178621042 @default.
- W2989742941 hasConceptScore W2989742941C187455244 @default.
- W2989742941 hasConceptScore W2989742941C19220575 @default.
- W2989742941 hasConceptScore W2989742941C24890656 @default.
- W2989742941 hasConceptScore W2989742941C2775997480 @default.
- W2989742941 hasConceptScore W2989742941C33923547 @default.
- W2989742941 hasConceptScore W2989742941C41008148 @default.
- W2989742941 hasConceptScore W2989742941C50820777 @default.
- W2989742941 hasConceptScore W2989742941C63584917 @default.
- W2989742941 hasConceptScore W2989742941C77553402 @default.
- W2989742941 hasConceptScore W2989742941C78458016 @default.
- W2989742941 hasConceptScore W2989742941C86803240 @default.
- W2989742941 hasLocation W29897429411 @default.
- W2989742941 hasOpenAccess W2989742941 @default.
- W2989742941 hasPrimaryLocation W29897429411 @default.
- W2989742941 hasRelatedWork W1789493849 @default.
- W2989742941 hasRelatedWork W2401469347 @default.
- W2989742941 hasRelatedWork W2562789087 @default.
- W2989742941 hasRelatedWork W2770219927 @default.