Matches in SemOpenAlex for { <https://semopenalex.org/work/W2962746230> ?p ?o ?g. }
- W2962746230 endingPage "20" @default.
- W2962746230 startingPage "10" @default.
- W2962746230 abstract "Many machine learning problems can be reduced to learning a low-rank positive semidefinite matrix (denoted as Z ), which encounters semidefinite program (SDP). Existing SDP solvers are often expensive for large-scale learning. In order to avoid directly solving SDP, some works convert SDP into a nonconvex program by factorizing Z as XX ⊤ . However, this would bring higher-order nonlinearity, resulting in scarcity of structure in subsequent optimization. In this paper, we propose a novel surrogate for SDP-based learning, in which the structure of subproblem is exploited. More specifically, we surrogate unconstrained SDP by a biconvex problem, through factorizing Z as XY ⊤ and using a quadratic penalty to penalize the difference of X and Y , in which the resultant subproblems are convex. Furthermore, we provide a theory bound for the associated penalty parameter under the assumption that the objective function is Lipschitz-smooth, such that the proposed surrogate will solve the original SDP when the penalty parameter is larger than this bound. Experiments on three SDP-based machine learning problems demonstrate that the proposed algorithm is as accurate as the state-of-the-art, but is faster on large-scale learning." @default.
- W2962746230 created "2019-07-30" @default.
- W2962746230 creator A5003334032 @default.
- W2962746230 creator A5081057351 @default.
- W2962746230 date "2019-11-01" @default.
- W2962746230 modified "2023-10-17" @default.
- W2962746230 title "A new optimization in SDP-based learning" @default.
- W2962746230 cites W1810986334 @default.
- W2962746230 cites W1830256639 @default.
- W2962746230 cites W1964590153 @default.
- W2962746230 cites W1987083649 @default.
- W2962746230 cites W2017588182 @default.
- W2962746230 cites W2019713176 @default.
- W2962746230 cites W2030161963 @default.
- W2962746230 cites W2033403400 @default.
- W2962746230 cites W2109904757 @default.
- W2962746230 cites W2117154319 @default.
- W2962746230 cites W2117154949 @default.
- W2962746230 cites W2117198760 @default.
- W2962746230 cites W2122090912 @default.
- W2962746230 cites W2130264611 @default.
- W2962746230 cites W2143075842 @default.
- W2962746230 cites W2146931605 @default.
- W2962746230 cites W2159514083 @default.
- W2962746230 cites W2183379929 @default.
- W2962746230 cites W2184634347 @default.
- W2962746230 cites W2285373897 @default.
- W2962746230 cites W2574720951 @default.
- W2962746230 cites W2615253071 @default.
- W2962746230 cites W2740578922 @default.
- W2962746230 cites W2751083812 @default.
- W2962746230 cites W2950082039 @default.
- W2962746230 cites W2950383522 @default.
- W2962746230 cites W2952975724 @default.
- W2962746230 cites W2963709905 @default.
- W2962746230 cites W2907452271 @default.
- W2962746230 doi "https://doi.org/10.1016/j.neucom.2019.06.058" @default.
- W2962746230 hasPublicationYear "2019" @default.
- W2962746230 type Work @default.
- W2962746230 sameAs 2962746230 @default.
- W2962746230 citedByCount "2" @default.
- W2962746230 countsByYear W29627462302020 @default.
- W2962746230 countsByYear W29627462302022 @default.
- W2962746230 crossrefType "journal-article" @default.
- W2962746230 hasAuthorship W2962746230A5003334032 @default.
- W2962746230 hasAuthorship W2962746230A5081057351 @default.
- W2962746230 hasConcept C101901036 @default.
- W2962746230 hasConcept C114614502 @default.
- W2962746230 hasConcept C121332964 @default.
- W2962746230 hasConcept C126255220 @default.
- W2962746230 hasConcept C129844170 @default.
- W2962746230 hasConcept C134306372 @default.
- W2962746230 hasConcept C14036430 @default.
- W2962746230 hasConcept C158693339 @default.
- W2962746230 hasConcept C164226766 @default.
- W2962746230 hasConcept C22324862 @default.
- W2962746230 hasConcept C2524010 @default.
- W2962746230 hasConcept C33923547 @default.
- W2962746230 hasConcept C41008148 @default.
- W2962746230 hasConcept C49712288 @default.
- W2962746230 hasConcept C62520636 @default.
- W2962746230 hasConcept C78458016 @default.
- W2962746230 hasConcept C86803240 @default.
- W2962746230 hasConceptScore W2962746230C101901036 @default.
- W2962746230 hasConceptScore W2962746230C114614502 @default.
- W2962746230 hasConceptScore W2962746230C121332964 @default.
- W2962746230 hasConceptScore W2962746230C126255220 @default.
- W2962746230 hasConceptScore W2962746230C129844170 @default.
- W2962746230 hasConceptScore W2962746230C134306372 @default.
- W2962746230 hasConceptScore W2962746230C14036430 @default.
- W2962746230 hasConceptScore W2962746230C158693339 @default.
- W2962746230 hasConceptScore W2962746230C164226766 @default.
- W2962746230 hasConceptScore W2962746230C22324862 @default.
- W2962746230 hasConceptScore W2962746230C2524010 @default.
- W2962746230 hasConceptScore W2962746230C33923547 @default.
- W2962746230 hasConceptScore W2962746230C41008148 @default.
- W2962746230 hasConceptScore W2962746230C49712288 @default.
- W2962746230 hasConceptScore W2962746230C62520636 @default.
- W2962746230 hasConceptScore W2962746230C78458016 @default.
- W2962746230 hasConceptScore W2962746230C86803240 @default.
- W2962746230 hasFunder F4320321001 @default.
- W2962746230 hasLocation W29627462301 @default.
- W2962746230 hasOpenAccess W2962746230 @default.
- W2962746230 hasPrimaryLocation W29627462301 @default.
- W2962746230 hasRelatedWork W1564910180 @default.
- W2962746230 hasRelatedWork W1996702873 @default.
- W2962746230 hasRelatedWork W2009581913 @default.
- W2962746230 hasRelatedWork W2027184514 @default.
- W2962746230 hasRelatedWork W2054129179 @default.
- W2962746230 hasRelatedWork W2064078067 @default.
- W2962746230 hasRelatedWork W2209046899 @default.
- W2962746230 hasRelatedWork W2270311210 @default.
- W2962746230 hasRelatedWork W2366968932 @default.
- W2962746230 hasRelatedWork W2586827781 @default.
- W2962746230 hasRelatedWork W2736546756 @default.
- W2962746230 hasRelatedWork W2791653687 @default.
- W2962746230 hasRelatedWork W2904185888 @default.
- W2962746230 hasRelatedWork W2951431231 @default.