FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W2948069823> ?p ?o ?g. }

• W2948069823 abstract "Traditionally in regression one minimizes the number of fitting parameters or uses smoothing/regularization to trade training (TE) and generalization error (GE). Driving TE to zero by increasing fitting degrees of freedom (dof) is expected to increase GE. However big-data approaches, including deep nets, seem to over-parametrize and send TE to zero (data interpolation) without impacting GE. Overparametrization has the benefit that global minima of the empirical loss function proliferate and become easier to find. These phenomena have drawn theoretical attention. and classification algorithms have been shown that interpolate data but also generalize optimally. An interesting related phenomenon has been noted: the existence of non-monotonic risk curves, with a peak in GE with increasing dof. It was suggested that this peak separates a regime from a regime where over-parametrization improves performance. Similar over-fitting peaks were reported previously (statistical physics approach to learning) and attributed to increased fitting model flexibility. We introduce a generative and fitting model pair (Misparametrized Sparse Regression or MiSpaR) and show that the overfitting peak can be dissociated from the point at which the fitting function gains enough dof's to match the data generative model and thus provides good generalization. This complicates the interpretation of overfitting peaks as separating a classical from a modern regime. Data interpolation itself cannot guarantee good generalization: we need to study the interpolation with different penalty terms. We present analytical formulae for GE curves for MiSpaR with $l_2$ and $l_1$ penalties, in the interpolating limit $lambdarightarrow 0$.These risk curves exhibit important differences and help elucidate the underlying phenomena." @default.
• W2948069823 created "2019-06-14" @default.
• W2948069823 creator A5037732748 @default.
• W2948069823 date "2019-06-09" @default.
• W2948069823 modified "2023-09-27" @default.
• W2948069823 title "Understanding overfitting peaks in generalization error: Analytical risk curves for l 2 and l 1 penalized interpolation." @default.
• W2948069823 cites W121410702 @default.
• W2948069823 cites W1981051810 @default.
• W2948069823 cites W2019705392 @default.
• W2948069823 cites W2046658845 @default.
• W2948069823 cites W2060581589 @default.
• W2948069823 cites W2082029531 @default.
• W2948069823 cites W2163916252 @default.
• W2948069823 cites W2259536334 @default.
• W2948069823 cites W2487770199 @default.
• W2948069823 cites W2566505556 @default.
• W2948069823 cites W2763894180 @default.
• W2948069823 cites W2807842867 @default.
• W2948069823 cites W2907127169 @default.
• W2948069823 cites W2923764619 @default.
• W2948069823 cites W2959995783 @default.
• W2948069823 cites W2963527807 @default.
• W2948069823 cites W2964034630 @default.
• W2948069823 cites W2972810859 @default.
• W2948069823 cites W3105033759 @default.
• W2948069823 cites W3106421180 @default.
• W2948069823 cites W3121195152 @default.
• W2948069823 cites W3137695714 @default.
• W2948069823 cites W3141350557 @default.
• W2948069823 hasPublicationYear "2019" @default.
• W2948069823 type Work @default.
• W2948069823 sameAs 2948069823 @default.
• W2948069823 citedByCount "29" @default.
• W2948069823 countsByYear W29480698232019 @default.
• W2948069823 countsByYear W29480698232020 @default.
• W2948069823 countsByYear W29480698232021 @default.
• W2948069823 crossrefType "posted-content" @default.
• W2948069823 hasAuthorship W2948069823A5037732748 @default.
• W2948069823 hasConcept C104114177 @default.
• W2948069823 hasConcept C105795698 @default.
• W2948069823 hasConcept C11413529 @default.
• W2948069823 hasConcept C126255220 @default.
• W2948069823 hasConcept C134306372 @default.
• W2948069823 hasConcept C137800194 @default.
• W2948069823 hasConcept C14036430 @default.
• W2948069823 hasConcept C154945302 @default.
• W2948069823 hasConcept C177148314 @default.
• W2948069823 hasConcept C186633575 @default.
• W2948069823 hasConcept C21080849 @default.
• W2948069823 hasConcept C22019652 @default.
• W2948069823 hasConcept C2776135515 @default.
• W2948069823 hasConcept C28826006 @default.
• W2948069823 hasConcept C33923547 @default.
• W2948069823 hasConcept C3770464 @default.
• W2948069823 hasConcept C41008148 @default.
• W2948069823 hasConcept C50644808 @default.
• W2948069823 hasConcept C51820054 @default.
• W2948069823 hasConcept C78458016 @default.
• W2948069823 hasConcept C86803240 @default.
• W2948069823 hasConceptScore W2948069823C104114177 @default.
• W2948069823 hasConceptScore W2948069823C105795698 @default.
• W2948069823 hasConceptScore W2948069823C11413529 @default.
• W2948069823 hasConceptScore W2948069823C126255220 @default.
• W2948069823 hasConceptScore W2948069823C134306372 @default.
• W2948069823 hasConceptScore W2948069823C137800194 @default.
• W2948069823 hasConceptScore W2948069823C14036430 @default.
• W2948069823 hasConceptScore W2948069823C154945302 @default.
• W2948069823 hasConceptScore W2948069823C177148314 @default.
• W2948069823 hasConceptScore W2948069823C186633575 @default.
• W2948069823 hasConceptScore W2948069823C21080849 @default.
• W2948069823 hasConceptScore W2948069823C22019652 @default.
• W2948069823 hasConceptScore W2948069823C2776135515 @default.
• W2948069823 hasConceptScore W2948069823C28826006 @default.
• W2948069823 hasConceptScore W2948069823C33923547 @default.
• W2948069823 hasConceptScore W2948069823C3770464 @default.
• W2948069823 hasConceptScore W2948069823C41008148 @default.
• W2948069823 hasConceptScore W2948069823C50644808 @default.
• W2948069823 hasConceptScore W2948069823C51820054 @default.
• W2948069823 hasConceptScore W2948069823C78458016 @default.
• W2948069823 hasConceptScore W2948069823C86803240 @default.
• W2948069823 hasLocation W29480698231 @default.
• W2948069823 hasOpenAccess W2948069823 @default.
• W2948069823 hasPrimaryLocation W29480698231 @default.
• W2948069823 hasRelatedWork W1975846642 @default.
• W2948069823 hasRelatedWork W2144902422 @default.
• W2948069823 hasRelatedWork W2763894180 @default.
• W2948069823 hasRelatedWork W2809090039 @default.
• W2948069823 hasRelatedWork W2923764619 @default.
• W2948069823 hasRelatedWork W2959995783 @default.
• W2948069823 hasRelatedWork W2963038274 @default.
• W2948069823 hasRelatedWork W2963518130 @default.
• W2948069823 hasRelatedWork W2963742538 @default.
• W2948069823 hasRelatedWork W2964034630 @default.
• W2948069823 hasRelatedWork W2964161291 @default.
• W2948069823 hasRelatedWork W2967536008 @default.
• W2948069823 hasRelatedWork W2976666284 @default.
• W2948069823 hasRelatedWork W2989948753 @default.
• W2948069823 hasRelatedWork W2996603747 @default.
• W2948069823 hasRelatedWork W3014316192 @default.
• W2948069823 hasRelatedWork W3018252856 @default.

• next