Matches in SemOpenAlex for { <https://semopenalex.org/work/W3110088573> ?p ?o ?g. }
Showing items 1 to 96 of
96
with 100 items per page.
- W3110088573 abstract "Acquisition of data is a difficult task in most applications of Machine Learning (ML), and it is only natural that one hopes and expects lower populating risk (better performance) with increasing data points. It turns out, somewhat surprisingly, that this is not the case even for the most standard algorithms such as the Empirical Risk Minimizer (ERM). Non-monotonic behaviour of the risk and instability in training have manifested and appeared in the popular deep learning paradigm under the description of double descent. These problems not only highlight our lack of understanding of learning algorithms and generalization but rather render our efforts at data acquisition in vain. It is, therefore, crucial to pursue this concern and provide a characterization of such behaviour. In this paper, we derive the first consistent and risk-monotonic algorithms for a general statistical learning setting under weak assumptions, consequently resolving an open problem (Viering et al. 2019) on how to avoid non-monotonic behaviour of risk curves. Our algorithms make use of Distributionally Robust Optimization (DRO) -- a technique that has shown promise in other complications of deep learning such as adversarial training. Our work makes a significant contribution to the topic of risk-monotonicity, which may be key in resolving empirical phenomena such as double descent." @default.
- W3110088573 created "2020-12-07" @default.
- W3110088573 creator A5033128676 @default.
- W3110088573 creator A5075917792 @default.
- W3110088573 date "2020-11-28" @default.
- W3110088573 modified "2023-09-27" @default.
- W3110088573 title "Risk-Monotonicity via Distributional Robustness." @default.
- W3110088573 cites W1651573571 @default.
- W3110088573 cites W1693992350 @default.
- W3110088573 cites W1962364105 @default.
- W3110088573 cites W2029538739 @default.
- W3110088573 cites W2031355801 @default.
- W3110088573 cites W2047581872 @default.
- W3110088573 cites W2053801139 @default.
- W3110088573 cites W2328111639 @default.
- W3110088573 cites W2407624838 @default.
- W3110088573 cites W2531499355 @default.
- W3110088573 cites W2579923771 @default.
- W3110088573 cites W2944407464 @default.
- W3110088573 cites W2963450292 @default.
- W3110088573 cites W2965233673 @default.
- W3110088573 cites W2967536008 @default.
- W3110088573 cites W2968009374 @default.
- W3110088573 cites W2970728427 @default.
- W3110088573 cites W2970874350 @default.
- W3110088573 cites W2972810859 @default.
- W3110088573 cites W2996279083 @default.
- W3110088573 cites W2996360471 @default.
- W3110088573 cites W3009460743 @default.
- W3110088573 cites W3046835365 @default.
- W3110088573 cites W3102072887 @default.
- W3110088573 cites W3104479411 @default.
- W3110088573 cites W3105741936 @default.
- W3110088573 cites W3151137625 @default.
- W3110088573 cites W607505555 @default.
- W3110088573 cites W86235050 @default.
- W3110088573 hasPublicationYear "2020" @default.
- W3110088573 type Work @default.
- W3110088573 sameAs 3110088573 @default.
- W3110088573 citedByCount "0" @default.
- W3110088573 crossrefType "posted-content" @default.
- W3110088573 hasAuthorship W3110088573A5033128676 @default.
- W3110088573 hasAuthorship W3110088573A5075917792 @default.
- W3110088573 hasConcept C104317684 @default.
- W3110088573 hasConcept C107321475 @default.
- W3110088573 hasConcept C119857082 @default.
- W3110088573 hasConcept C134306372 @default.
- W3110088573 hasConcept C154945302 @default.
- W3110088573 hasConcept C177148314 @default.
- W3110088573 hasConcept C185592680 @default.
- W3110088573 hasConcept C33923547 @default.
- W3110088573 hasConcept C37736160 @default.
- W3110088573 hasConcept C41008148 @default.
- W3110088573 hasConcept C55493867 @default.
- W3110088573 hasConcept C63479239 @default.
- W3110088573 hasConcept C72169020 @default.
- W3110088573 hasConceptScore W3110088573C104317684 @default.
- W3110088573 hasConceptScore W3110088573C107321475 @default.
- W3110088573 hasConceptScore W3110088573C119857082 @default.
- W3110088573 hasConceptScore W3110088573C134306372 @default.
- W3110088573 hasConceptScore W3110088573C154945302 @default.
- W3110088573 hasConceptScore W3110088573C177148314 @default.
- W3110088573 hasConceptScore W3110088573C185592680 @default.
- W3110088573 hasConceptScore W3110088573C33923547 @default.
- W3110088573 hasConceptScore W3110088573C37736160 @default.
- W3110088573 hasConceptScore W3110088573C41008148 @default.
- W3110088573 hasConceptScore W3110088573C55493867 @default.
- W3110088573 hasConceptScore W3110088573C63479239 @default.
- W3110088573 hasConceptScore W3110088573C72169020 @default.
- W3110088573 hasLocation W31100885731 @default.
- W3110088573 hasOpenAccess W3110088573 @default.
- W3110088573 hasPrimaryLocation W31100885731 @default.
- W3110088573 hasRelatedWork W1483390608 @default.
- W3110088573 hasRelatedWork W1569657508 @default.
- W3110088573 hasRelatedWork W1931877416 @default.
- W3110088573 hasRelatedWork W1969146585 @default.
- W3110088573 hasRelatedWork W2167567739 @default.
- W3110088573 hasRelatedWork W2771004121 @default.
- W3110088573 hasRelatedWork W2775694264 @default.
- W3110088573 hasRelatedWork W2799385102 @default.
- W3110088573 hasRelatedWork W2890211540 @default.
- W3110088573 hasRelatedWork W2900971815 @default.
- W3110088573 hasRelatedWork W2952981590 @default.
- W3110088573 hasRelatedWork W2979895842 @default.
- W3110088573 hasRelatedWork W3011584947 @default.
- W3110088573 hasRelatedWork W3020682900 @default.
- W3110088573 hasRelatedWork W3094709024 @default.
- W3110088573 hasRelatedWork W3105793250 @default.
- W3110088573 hasRelatedWork W3113847213 @default.
- W3110088573 hasRelatedWork W3151269043 @default.
- W3110088573 hasRelatedWork W3207729118 @default.
- W3110088573 hasRelatedWork W3209838601 @default.
- W3110088573 isParatext "false" @default.
- W3110088573 isRetracted "false" @default.
- W3110088573 magId "3110088573" @default.
- W3110088573 workType "article" @default.