Zobrazeno 1 - 10
of 16
pro vyhledávání: '"Erik Kruus"'
Publikováno v:
Proceedings of the 28th ACM SIGKDD Conference on Knowledge Discovery and Data Mining.
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_________::a8782e4e17c853591694000186fa262c
https://doi.org/10.1145/3534678.3539075
https://doi.org/10.1145/3534678.3539075
We study few-shot debugging of transformer based natural language understanding models, using recently popularized test suites to not just diagnose but correct a problem. Given a few debugging examples of a certain phenomenon, and a held-out test set
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_dedup___::f7f91d7c72c10a10cc6c770648c71dd0
Publikováno v:
2021 IEEE/CVF International Conference on Computer Vision (ICCV).
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_________::e29535c98f21c6a26e3d758c85363970
https://doi.org/10.1109/iccv48922.2021.00740
https://doi.org/10.1109/iccv48922.2021.00740
Publikováno v:
Pattern Recognition. ICPR International Workshops and Challenges ISBN: 9783030687793
ICPR Workshops (6)
ICPR Workshops (6)
One major source of vulnerability of neural nets in classification tasks is from overparameterized fully connected layers near the end of the network. In this paper, we propose a new neighborhood preserving layer which can replace these fully connect
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_________::6c6ee59fc34a81aa00961c0c3014daea
https://doi.org/10.1007/978-3-030-68780-9_17
https://doi.org/10.1007/978-3-030-68780-9_17
Publikováno v:
EuroSys
Machine learning methods, such as SVM and neural networks, often improve their accuracy by using models with more parameters trained on large numbers of examples. Building such models on a single machine is often impractical because of the large amou
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_________::bf63b985fe351ac906a0f37af6108c8f
https://doi.org/10.1145/2741948.2741965
https://doi.org/10.1145/2741948.2741965
Publikováno v:
Nucleic Acids Research
Protein backbones have characteristic secondary structures, including alpha-helices and beta-sheets. Which structure is adopted locally is strongly biased by the local amino acid sequence of the protein. Accurate (probabilistic) mappings from sequenc
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_dedup___::1baeb1bbd53a814ff4a9b03a0c20d751
https://doi.org/10.1093/nar/gki842
https://doi.org/10.1093/nar/gki842
Publikováno v:
Langmuir. 14:6579-6587
Layered materials composed of silver−alkanethiolate units are characterized and compared to self-assembled alkanethiolate monolayers and thiolate-capped nanoparticles. A comprehensive infrared spectroscopy study is presented for silver−alkanethio
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_________::97ba530c3609df4d4452b914e5263824
https://doi.org/10.1021/la980718g
https://doi.org/10.1021/la980718g
Publikováno v:
Canadian Journal of Chemistry. 76:1654-1659
available
unlimited
public
unlimited
public
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_dedup___::756248cd348bc5b16586e71a089944c2
https://doi.org/10.1139/cjc-76-11-1654
https://doi.org/10.1139/cjc-76-11-1654
Publikováno v:
Langmuir. 13:5335-5340
The effect of immersion time on the ordering properties of self-assembled docosanethiolate (C22) monolayers has been studied by reflection/absorption infrared spectroscopy (RAIRS). Starting with a chemically cleaned polycrystalline gold surface, we h
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_________::7c6bf960f66970e07c20c16a217f5a17
https://doi.org/10.1021/la970052a
https://doi.org/10.1021/la970052a