Zobrazeno 1 - 10
of 10
pro vyhledávání: '"Urun Dogan"'
Publikováno v:
PLoS ONE, Vol 12, Iss 6, p e0178161 (2017)
Training of one-vs.-rest SVMs can be parallelized over the number of classes in a straight forward way. Given enough computational resources, one-vs.-rest SVMs can thus be trained on data involving a large number of classes. The same cannot be stated
Externí odkaz:
https://doaj.org/article/061471f88a47404785ce126bd0a1fd7d
In this paper, we focus on unsupervised representation learning for clustering of images. Recent advances in deep clustering and unsupervised representation learning are based on the idea that different views of an input image (generated through data
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_dedup___::ffd5f967df396ea18aab0ec849a08b90
Publikováno v:
Computer Vision – ECCV 2020 ISBN: 9783030585259
ECCV (29)
ECCV (29)
Curriculum learning can improve neural network training by guiding the optimization to desirable optima. We propose a novel curriculum learning approach for image classification that adapts the loss function by changing the label representation.
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_________::efef610c62db19dd0888e3e29983516b
https://doi.org/10.1007/978-3-030-58526-6_11
https://doi.org/10.1007/978-3-030-58526-6_11
Publikováno v:
IJCNN
Recent advances in deep clustering and unsupervised representation learning are based on the idea that different views of an input image (generated through data augmentation techniques) must either be closer in the representation space, or have a sim
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_dedup___::a97a5551e524e49e23bccc0dcfa5cb94
Publikováno v:
PLoS ONE
PLoS ONE, Vol 12, Iss 6, p e0178161 (2017)
PLoS ONE, Vol 12, Iss 6, p e0178161 (2017)
Training of one-vs.-rest SVMs can be parallelized over the number of classes in a straight forward way. Given enough computational resources, one-vs.-rest SVMs can thus be trained on data involving a large number of classes. The same cannot be stated
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_dedup___::baf810eb2bcf973ecc422717ecaa908f
http://europepmc.org/articles/PMC5453486
http://europepmc.org/articles/PMC5453486
In this paper, we study data-dependent generalization error bounds that exhibit a mild dependency on the number of classes, making them suitable for multi-class learning with a large number of label classes. The bounds generally hold for empirical mu
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_dedup___::01149f23661a079c9599d835ba8c1778
Publikováno v:
Machine Learning and Knowledge Discovery in Databases ISBN: 9783642334597
ECML/PKDD (1)
ECML/PKDD (1)
A generic way to extend generalization bounds for binary large-margin classifiers to large-margin multi-category classifiers is presented. The simple proceeding leads to surprisingly tight bounds showing the same $\tilde{O}(d^2)$ scaling in the numbe
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_________::1f85780b32929f9e33ab4881ae6ab73a
https://doi.org/10.1007/978-3-642-33460-3_13
https://doi.org/10.1007/978-3-642-33460-3_13
Publikováno v:
Lecture Notes in Computer Science ISBN: 9783642327162
DAGM/OAGM Symposium
DAGM/OAGM Symposium
Stability selection [9] is a general principle for performing feature selection. It functions as a meta-layer on top of a “baseline” feature selection method, and consists in repeatedly applying the baseline to random data subsamples of half-size
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_________::b74409bc5df44d0c478465d91fc0987d
https://doi.org/10.1007/978-3-642-32717-9_26
https://doi.org/10.1007/978-3-642-32717-9_26
Publikováno v:
ROBIO
In the presented work we compare machine learning techniques in the context of lane change behavior performed by humans in a semi-naturalistic simulated environment. We evaluate different learning approaches using differing feature combinations in or
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_________::d807be76febc948fb1ea8c662b30e549
https://doi.org/10.1109/robio.2011.6181557
https://doi.org/10.1109/robio.2011.6181557
Publikováno v:
ITSC
The presented work formulates an framework in which early prediction of drivers lane change behavior is realized. We aim to build a representation of drivers lane change behavior in order to recognize and to predict driver's intentions as a first ste
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_________::2281579b58d5821ffee4dbb405081967
https://doi.org/10.1109/itsc.2008.4732700
https://doi.org/10.1109/itsc.2008.4732700