FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W2804540002> ?p ?o ?g. }

• W2804540002 abstract "Learning with a {it convex loss} function has been a dominating paradigm for many years. It remains an interesting question how non-convex loss functions help improve the generalization of learning with broad applicability. In this paper, we study a family of objective functions formed by truncating traditional loss functions, which is applicable to both shallow learning and deep learning. Truncating loss functions has potential to be less vulnerable and more robust to large noise in observations that could be adversarial. More importantly, it is a generic technique without assuming the knowledge of noise distribution. To justify non-convex learning with truncated losses, we establish excess risk bounds of empirical risk minimization based on truncated losses for heavy-tailed output, and statistical error of an approximate stationary point found by stochastic gradient descent (SGD) method. Our experiments for shallow and deep learning for regression with outliers, corrupted data and heavy-tailed noise further justify the proposed method." @default.
• W2804540002 created "2018-06-01" @default.
• W2804540002 creator A5002854172 @default.
• W2804540002 creator A5018343158 @default.
• W2804540002 creator A5023288846 @default.
• W2804540002 creator A5069394608 @default.
• W2804540002 creator A5070881368 @default.
• W2804540002 creator A5088141052 @default.
• W2804540002 date "2018-05-21" @default.
• W2804540002 modified "2023-09-27" @default.
• W2804540002 title "Learning with Non-Convex Truncated Losses by SGD" @default.
• W2804540002 cites W114517082 @default.
• W2804540002 cites W149129625 @default.
• W2804540002 cites W1984332158 @default.
• W2804540002 cites W2026523206 @default.
• W2804540002 cites W2034365297 @default.
• W2804540002 cites W2047278710 @default.
• W2804540002 cites W2057690238 @default.
• W2804540002 cites W2080745194 @default.
• W2804540002 cites W2089493818 @default.
• W2804540002 cites W2113118796 @default.
• W2804540002 cites W2130315424 @default.
• W2804540002 cites W2345885508 @default.
• W2804540002 cites W2527354364 @default.
• W2804540002 cites W2597367899 @default.
• W2804540002 cites W2786313301 @default.
• W2804540002 cites W2798818710 @default.
• W2804540002 cites W2949125143 @default.
• W2804540002 cites W2949521215 @default.
• W2804540002 cites W2950468429 @default.
• W2804540002 cites W2951396696 @default.
• W2804540002 cites W2962896218 @default.
• W2804540002 cites W2963449842 @default.
• W2804540002 cites W2963517436 @default.
• W2804540002 cites W2963572540 @default.
• W2804540002 cites W2963927498 @default.
• W2804540002 cites W2963991711 @default.
• W2804540002 cites W3102654945 @default.
• W2804540002 cites W3104898220 @default.
• W2804540002 cites W3187217851 @default.
• W2804540002 cites W648260396 @default.
• W2804540002 cites W2019781618 @default.
• W2804540002 hasPublicationYear "2018" @default.
• W2804540002 type Work @default.
• W2804540002 sameAs 2804540002 @default.
• W2804540002 citedByCount "3" @default.
• W2804540002 countsByYear W28045400022018 @default.
• W2804540002 countsByYear W28045400022019 @default.
• W2804540002 crossrefType "posted-content" @default.
• W2804540002 hasAuthorship W2804540002A5002854172 @default.
• W2804540002 hasAuthorship W2804540002A5018343158 @default.
• W2804540002 hasAuthorship W2804540002A5023288846 @default.
• W2804540002 hasAuthorship W2804540002A5069394608 @default.
• W2804540002 hasAuthorship W2804540002A5070881368 @default.
• W2804540002 hasAuthorship W2804540002A5088141052 @default.
• W2804540002 hasConcept C107321475 @default.
• W2804540002 hasConcept C112680207 @default.
• W2804540002 hasConcept C115961682 @default.
• W2804540002 hasConcept C126255220 @default.
• W2804540002 hasConcept C134306372 @default.
• W2804540002 hasConcept C14036430 @default.
• W2804540002 hasConcept C145446738 @default.
• W2804540002 hasConcept C147764199 @default.
• W2804540002 hasConcept C154945302 @default.
• W2804540002 hasConcept C157972887 @default.
• W2804540002 hasConcept C177148314 @default.
• W2804540002 hasConcept C206688291 @default.
• W2804540002 hasConcept C2524010 @default.
• W2804540002 hasConcept C28826006 @default.
• W2804540002 hasConcept C33923547 @default.
• W2804540002 hasConcept C41008148 @default.
• W2804540002 hasConcept C50644808 @default.
• W2804540002 hasConcept C78458016 @default.
• W2804540002 hasConcept C79337645 @default.
• W2804540002 hasConcept C86803240 @default.
• W2804540002 hasConcept C99498987 @default.
• W2804540002 hasConceptScore W2804540002C107321475 @default.
• W2804540002 hasConceptScore W2804540002C112680207 @default.
• W2804540002 hasConceptScore W2804540002C115961682 @default.
• W2804540002 hasConceptScore W2804540002C126255220 @default.
• W2804540002 hasConceptScore W2804540002C134306372 @default.
• W2804540002 hasConceptScore W2804540002C14036430 @default.
• W2804540002 hasConceptScore W2804540002C145446738 @default.
• W2804540002 hasConceptScore W2804540002C147764199 @default.
• W2804540002 hasConceptScore W2804540002C154945302 @default.
• W2804540002 hasConceptScore W2804540002C157972887 @default.
• W2804540002 hasConceptScore W2804540002C177148314 @default.
• W2804540002 hasConceptScore W2804540002C206688291 @default.
• W2804540002 hasConceptScore W2804540002C2524010 @default.
• W2804540002 hasConceptScore W2804540002C28826006 @default.
• W2804540002 hasConceptScore W2804540002C33923547 @default.
• W2804540002 hasConceptScore W2804540002C41008148 @default.
• W2804540002 hasConceptScore W2804540002C50644808 @default.
• W2804540002 hasConceptScore W2804540002C78458016 @default.
• W2804540002 hasConceptScore W2804540002C79337645 @default.
• W2804540002 hasConceptScore W2804540002C86803240 @default.
• W2804540002 hasConceptScore W2804540002C99498987 @default.
• W2804540002 hasLocation W28045400021 @default.
• W2804540002 hasOpenAccess W2804540002 @default.
• W2804540002 hasPrimaryLocation W28045400021 @default.

• next