Matches in SemOpenAlex for { <https://semopenalex.org/work/W3201059461> ?p ?o ?g. }
- W3201059461 abstract "We propose a deep distributionally robust learning framework for calibrated uncertainties under domain shifts. We consider cases where the source (training) distribution differs significantly from the target (test) distribution. In addition to the standard class predictor, our framework contains a binary domain classifier which estimates the density ratio between the source and target domains. We incorporate both with neural networks and train them end-to-end. The framework is demonstrated to generate calibrated uncertainties that benefit many downstream tasks, including unsupervised domain adaptation (UDA) and semi-supervised learning (SSL) where methods such as self-training and FixMatch use uncertainties to select confident pseudo-labels. Our experiments show that the introduction of DRL to these methods leads to significant improvements in cross-domain performance. We also demonstrate that the produced density ratio estimates show agreement with the human selection frequencies, suggesting a match with the human perceived uncertainties. The source code of this work will be made publicly available." @default.
- W3201059461 created "2021-09-27" @default.
- W3201059461 creator A5000220618 @default.
- W3201059461 creator A5014498545 @default.
- W3201059461 creator A5054586129 @default.
- W3201059461 creator A5077576020 @default.
- W3201059461 creator A5085826758 @default.
- W3201059461 creator A5087158377 @default.
- W3201059461 date "2020-10-08" @default.
- W3201059461 modified "2023-09-27" @default.
- W3201059461 title "Deep Distributionally Robust Learning for Calibrated Uncertainties under Domain Shift" @default.
- W3201059461 cites W1484551447 @default.
- W3201059461 cites W1577269164 @default.
- W3201059461 cites W1590693676 @default.
- W3201059461 cites W1618905105 @default.
- W3201059461 cites W1686810756 @default.
- W3201059461 cites W1722318740 @default.
- W3201059461 cites W1809653203 @default.
- W3201059461 cites W1966026565 @default.
- W3201059461 cites W2023943903 @default.
- W3201059461 cites W2034368206 @default.
- W3201059461 cites W2064447488 @default.
- W3201059461 cites W2073241381 @default.
- W3201059461 cites W2080441299 @default.
- W3201059461 cites W2108018401 @default.
- W3201059461 cites W2108598243 @default.
- W3201059461 cites W2112796928 @default.
- W3201059461 cites W2115403315 @default.
- W3201059461 cites W2116200501 @default.
- W3201059461 cites W2118858186 @default.
- W3201059461 cites W2144193737 @default.
- W3201059461 cites W2159291411 @default.
- W3201059461 cites W2163605009 @default.
- W3201059461 cites W2164411961 @default.
- W3201059461 cites W2167433878 @default.
- W3201059461 cites W2183341477 @default.
- W3201059461 cites W2194775991 @default.
- W3201059461 cites W2212660284 @default.
- W3201059461 cites W2254249950 @default.
- W3201059461 cites W2335728318 @default.
- W3201059461 cites W2413199960 @default.
- W3201059461 cites W2549334782 @default.
- W3201059461 cites W2592505114 @default.
- W3201059461 cites W2600383743 @default.
- W3201059461 cites W2605488490 @default.
- W3201059461 cites W2766897166 @default.
- W3201059461 cites W2767075075 @default.
- W3201059461 cites W2778783904 @default.
- W3201059461 cites W2788557041 @default.
- W3201059461 cites W2795155917 @default.
- W3201059461 cites W2804697534 @default.
- W3201059461 cites W2884771968 @default.
- W3201059461 cites W2895281799 @default.
- W3201059461 cites W2895779420 @default.
- W3201059461 cites W2904967782 @default.
- W3201059461 cites W2912168444 @default.
- W3201059461 cites W2920306970 @default.
- W3201059461 cites W2945880513 @default.
- W3201059461 cites W2946063682 @default.
- W3201059461 cites W2949987290 @default.
- W3201059461 cites W2950300355 @default.
- W3201059461 cites W2951004968 @default.
- W3201059461 cites W2951111165 @default.
- W3201059461 cites W2951266961 @default.
- W3201059461 cites W2951626887 @default.
- W3201059461 cites W2962687275 @default.
- W3201059461 cites W2962739340 @default.
- W3201059461 cites W2962900737 @default.
- W3201059461 cites W2962915600 @default.
- W3201059461 cites W2963081736 @default.
- W3201059461 cites W2963238274 @default.
- W3201059461 cites W2963275094 @default.
- W3201059461 cites W2964059111 @default.
- W3201059461 cites W2964139811 @default.
- W3201059461 cites W2964159205 @default.
- W3201059461 cites W2964212410 @default.
- W3201059461 cites W2964288524 @default.
- W3201059461 cites W2964330179 @default.
- W3201059461 cites W2969338701 @default.
- W3201059461 cites W2970780203 @default.
- W3201059461 cites W2970859221 @default.
- W3201059461 cites W2971130081 @default.
- W3201059461 cites W2985406498 @default.
- W3201059461 cites W2985564719 @default.
- W3201059461 cites W2989700832 @default.
- W3201059461 cites W3034481430 @default.
- W3201059461 cites W3034932471 @default.
- W3201059461 cites W3035000591 @default.
- W3201059461 cites W3035675766 @default.
- W3201059461 cites W3037590790 @default.
- W3201059461 cites W3037775931 @default.
- W3201059461 cites W3042045432 @default.
- W3201059461 cites W3099570996 @default.
- W3201059461 cites W3106031957 @default.
- W3201059461 cites W3107744858 @default.
- W3201059461 cites W3111862348 @default.
- W3201059461 cites W81851533 @default.
- W3201059461 cites W91088564 @default.
- W3201059461 hasPublicationYear "2020" @default.
- W3201059461 type Work @default.