Matches in SemOpenAlex for { <https://semopenalex.org/work/W3043972982> ?p ?o ?g. }
- W3043972982 abstract "Optimising a ranking-based metric, such as Average Precision (AP), is notoriously challenging due to the fact that it is non-differentiable, and hence cannot be optimised directly using gradient-descent methods. To this end, we introduce an objective that optimises instead a smoothed approximation of AP, coined Smooth-AP. Smooth-AP is a plug-and-play objective function that allows for end-to-end training of deep networks with a simple and elegant implementation. We also present an analysis for why directly optimising the ranking based metric of AP offers benefits over other deep metric learning losses. We apply Smooth-AP to standard retrieval benchmarks: Stanford Online products and VehicleID, and also evaluate on larger-scale datasets: INaturalist for fine-grained category retrieval, and VGGFace2 and IJB-C for face retrieval. In all cases, we improve the performance over the state-of-the-art, especially for larger-scale datasets, thus demonstrating the effectiveness and scalability of Smooth-AP to real-world scenarios." @default.
- W3043972982 created "2020-07-29" @default.
- W3043972982 creator A5039721396 @default.
- W3043972982 creator A5048400549 @default.
- W3043972982 creator A5057678172 @default.
- W3043972982 creator A5076097168 @default.
- W3043972982 date "2020-07-23" @default.
- W3043972982 modified "2023-09-27" @default.
- W3043972982 title "Smooth-AP: Smoothing the Path Towards Large-Scale Image Retrieval." @default.
- W3043972982 cites W1532325895 @default.
- W3043972982 cites W1556531089 @default.
- W3043972982 cites W1797268635 @default.
- W3043972982 cites W1974160425 @default.
- W3043972982 cites W1975517671 @default.
- W3043972982 cites W1984309565 @default.
- W3043972982 cites W2001832483 @default.
- W3043972982 cites W2012592962 @default.
- W3043972982 cites W2023991840 @default.
- W3043972982 cites W204268067 @default.
- W3043972982 cites W2055667362 @default.
- W3043972982 cites W2059001985 @default.
- W3043972982 cites W2071027807 @default.
- W3043972982 cites W2103924867 @default.
- W3043972982 cites W2106053110 @default.
- W3043972982 cites W2108862644 @default.
- W3043972982 cites W2117539524 @default.
- W3043972982 cites W2124509324 @default.
- W3043972982 cites W2127176025 @default.
- W3043972982 cites W2128877075 @default.
- W3043972982 cites W2130556178 @default.
- W3043972982 cites W2131846894 @default.
- W3043972982 cites W2141362318 @default.
- W3043972982 cites W2157364932 @default.
- W3043972982 cites W2158139921 @default.
- W3043972982 cites W2179042386 @default.
- W3043972982 cites W2194775991 @default.
- W3043972982 cites W2340690086 @default.
- W3043972982 cites W2514087538 @default.
- W3043972982 cites W2527197300 @default.
- W3043972982 cites W2547446130 @default.
- W3043972982 cites W2549858646 @default.
- W3043972982 cites W2605102252 @default.
- W3043972982 cites W2606611007 @default.
- W3043972982 cites W2613997951 @default.
- W3043972982 cites W2740643007 @default.
- W3043972982 cites W2779066530 @default.
- W3043972982 cites W2797977484 @default.
- W3043972982 cites W2798303923 @default.
- W3043972982 cites W2798463715 @default.
- W3043972982 cites W2871667416 @default.
- W3043972982 cites W2885563408 @default.
- W3043972982 cites W2895347732 @default.
- W3043972982 cites W2921310091 @default.
- W3043972982 cites W2948077755 @default.
- W3043972982 cites W2948638722 @default.
- W3043972982 cites W2953271441 @default.
- W3043972982 cites W2962723992 @default.
- W3043972982 cites W2963026686 @default.
- W3043972982 cites W2963125676 @default.
- W3043972982 cites W2963284197 @default.
- W3043972982 cites W2963305974 @default.
- W3043972982 cites W2963350250 @default.
- W3043972982 cites W2963420686 @default.
- W3043972982 cites W2963466847 @default.
- W3043972982 cites W2963839617 @default.
- W3043972982 cites W2963988212 @default.
- W3043972982 cites W2964049460 @default.
- W3043972982 cites W2964121744 @default.
- W3043972982 cites W2964271799 @default.
- W3043972982 cites W2969875432 @default.
- W3043972982 cites W2969985801 @default.
- W3043972982 cites W2973134629 @default.
- W3043972982 cites W2982693525 @default.
- W3043972982 cites W2984884190 @default.
- W3043972982 cites W2990519439 @default.
- W3043972982 cites W2991234496 @default.
- W3043972982 cites W2991645104 @default.
- W3043972982 cites W2992827186 @default.
- W3043972982 cites W3023644007 @default.
- W3043972982 cites W3034941949 @default.
- W3043972982 cites W3035014997 @default.
- W3043972982 cites W3116566453 @default.
- W3043972982 hasPublicationYear "2020" @default.
- W3043972982 type Work @default.
- W3043972982 sameAs 3043972982 @default.
- W3043972982 citedByCount "2" @default.
- W3043972982 countsByYear W30439729822020 @default.
- W3043972982 countsByYear W30439729822021 @default.
- W3043972982 crossrefType "posted-content" @default.
- W3043972982 hasAuthorship W3043972982A5039721396 @default.
- W3043972982 hasAuthorship W3043972982A5048400549 @default.
- W3043972982 hasAuthorship W3043972982A5057678172 @default.
- W3043972982 hasAuthorship W3043972982A5076097168 @default.
- W3043972982 hasConcept C108583219 @default.
- W3043972982 hasConcept C115961682 @default.
- W3043972982 hasConcept C119857082 @default.
- W3043972982 hasConcept C121332964 @default.
- W3043972982 hasConcept C124101348 @default.
- W3043972982 hasConcept C134306372 @default.
- W3043972982 hasConcept C153258448 @default.