Zobrazeno 1 - 10
of 2 699
pro vyhledávání: '"Ye,Bin"'
Vision language models (VLMs) have shown promising reasoning capabilities across various benchmarks; however, our understanding of their visual perception remains limited. In this work, we propose an eye examination process to investigate how a VLM p
Externí odkaz:
http://arxiv.org/abs/2409.14759
Vision language models (VLMs) perceive the world through a combination of a visual encoder and a large language model (LLM). The visual encoder, pre-trained on large-scale vision-text datasets, provides zero-shot generalization to visual data, and th
Externí odkaz:
http://arxiv.org/abs/2407.13442
Autor:
YE Bin, WANG Changcheng
Publikováno v:
Zhongguo linchuang yanjiu, Vol 37, Iss 9, Pp 1347-1352 (2024)
Objective To analyze the relationship between immune-related markers ［A proliferation-inducing ligand/tumor necrosis factor superfamily 13(APRIL/TNFSF13), B lymphocyte activating factor (BAFF) and lymphocyte activating gene 3 (LAG-3)］ and clinico
Externí odkaz:
https://doaj.org/article/f5cc1e6071b742c5880e0131fd86eaa8
Data imbalance in training data often leads to biased predictions from trained models, which in turn causes ethical and social issues. A straightforward solution is to carefully curate training data, but given the enormous scale of modern neural netw
Externí odkaz:
http://arxiv.org/abs/2308.00994
We propose TextManiA, a text-driven manifold augmentation method that semantically enriches visual feature spaces, regardless of class distribution. TextManiA augments visual data with intra-class semantic perturbation by exploiting easy-to-understan
Externí odkaz:
http://arxiv.org/abs/2307.14611
Publikováno v:
Archives of Metallurgy and Materials, Vol vol. 69, Iss No 2, Pp 459-463 (2024)
Semiconducting GaN can realize high performance electronic and power devices owing to its high electron mobility and thermal conductivity where good metal-semiconductor contact is prerequisite. In this work, using thermal atomic layer deposition (ALD
Externí odkaz:
https://doaj.org/article/c93f3d1f1ab74fd2a1f695e3246c190d
We address a weakly-supervised low-shot instance segmentation, an annotation-efficient training method to deal with novel classes effectively. Since it is an under-explored problem, we first investigate the difficulty of the problem and identify the
Externí odkaz:
http://arxiv.org/abs/2302.09765
Publikováno v:
Food Science & Nutrition. Oct2024, Vol. 12 Issue 10, p7481-7491. 11p.
Autor:
Kim, Ye Bin1 (AUTHOR) zxc5620@gachon.ac.kr, Park, Seon Young2 (AUTHOR) lovesun139@snu.ac.kr, Jeon, Hye Jin3 (AUTHOR) jhj1125@knu.ac.kr, Kim, Bumkeun3 (AUTHOR) aam_kim@knu.ac.kr, Kwon, Mun-Gyeong4 (AUTHOR) mgkwon@korea.kr, Kim, Su-Mi4 (AUTHOR) sumikim@korea.kr, Han, Jee Eun3 (AUTHOR) jehan@knu.ac.kr, Kim, Ji Hyung1 (AUTHOR) jehan@knu.ac.kr
Publikováno v:
Animals (2076-2615). Oct2024, Vol. 14 Issue 19, p2788. 16p.
Autor:
Shin, Ye Bin1 (AUTHOR), Kim, Jin Hwan1 (AUTHOR), Kwon, Min Kyeong1 (AUTHOR), Myung, Jin Hyuk1 (AUTHOR), Lee, Dong Geon1 (AUTHOR), Jin, Sung Giu2 (AUTHOR), Kang, Myung Joo1 (AUTHOR) kangmj@dankook.ac.kr, Choi, Yong Seok1 (AUTHOR) analysc@dankook.ac.kr
Publikováno v:
PLoS ONE. 9/6/2024, Vol. 19 Issue 9, p1-16. 16p.