Zobrazeno 1 - 10
of 158
pro vyhledávání: '"Yoon, Jongmin"'
Autor:
Moon, Taehong, Choi, Moonseok, Yun, EungGu, Yoon, Jongmin, Lee, Gayoung, Cho, Jaewoong, Lee, Juho
Diffusion models have shown remarkable performance in generation problems over various domains including images, videos, text, and audio. A practical bottleneck of diffusion models is their sampling speed, due to the repeated evaluation of score esti
Externí odkaz:
http://arxiv.org/abs/2408.05927
Publikováno v:
ICLR 2024
Deep Ensemble (DE) approach is a straightforward technique used to enhance the performance of deep neural networks by training them from different initial points, converging towards various local optima. However, a limitation of this methodology lies
Externí odkaz:
http://arxiv.org/abs/2404.15814
Autor:
Yoon, Jongmin, Lee, Juho
Straightening the probability flow of the continuous-time generative models, such as diffusion models or flow-based models, is the key to fast sampling through the numerical solvers, existing methods learn a linear path by directly generating the pro
Externí odkaz:
http://arxiv.org/abs/2402.06461
Autor:
Yoon, Jongmin, Song, Haengjin, Park, Ji Soo, Kim, Jeong Ho, Jun, Yearin, Gim, Sang-Ah, Hong, Changhee, An, Kyung Mi, Park, Joon-Tae, Lee, Jung Woo, Yoon, Hongchul, Kim, Yun Seok, Kim, Sang Geon
Publikováno v:
In Biomedicine & Pharmacotherapy December 2024 181
Deep ensembles excel in large-scale image classification tasks both in terms of prediction accuracy and calibration. Despite being simple to train, the computation and memory cost of deep ensembles limits their practicability. While some recent works
Externí odkaz:
http://arxiv.org/abs/2110.14149
Autor:
Yoon, Jongmin, Lee, Don-Gil, Song, Haengjin, Hong, Dahae, Park, Ji Soo, Hong, Changhee, An, Kyung Mi, Lee, Jung Woo, Park, Joon-Tae, Yoon, Hongchul, Tak, Jihoon, Kim, Sang Geon
Publikováno v:
In Biomedicine & Pharmacotherapy August 2024 177
While adversarial training is considered as a standard defense method against adversarial attacks for image classifiers, adversarial purification, which purifies attacked images into clean images with a standalone purification model, has shown promis
Externí odkaz:
http://arxiv.org/abs/2106.06041
Autor:
Seong, Daekyeong, Yoon, Jongmin, Choo, Gyo-Hwang, Chang, Dong Yeong, Yang, Geum-Hee, Lee, Dae Gyun
Publikováno v:
In Environmental Pollution 1 April 2024 346
Autor:
Yoon, Jongmin, Peskin, Michael E.
We compute fermion pair production cross sections in $e^+e^-$ annihilation in models of electroweak symmetry breaking in warped 5-dimensional space. Our analysis is based on a framework with no elementary scalars, only gauge bosons and fermions in th
Externí odkaz:
http://arxiv.org/abs/1811.07877
Autor:
Yoon, Jongmin, Peskin, Michael E.
Publikováno v:
Phys. Rev. D 100, 015001 (2019)
We analyze models of electroweak symmetry breaking in warped 5-dimensional space with gauge bosons and fermions in the bulk. The Higgs boson is identified with the 5th component of a gauge field. We dynamically generate the Higgs potential using a co
Externí odkaz:
http://arxiv.org/abs/1810.12352