Zobrazeno 1 - 10
of 114 766
pro vyhledávání: '"Sugiyama, A"'
Recent advances in fine-tuning Vision-Language Models (VLMs) have witnessed the success of prompt tuning and adapter tuning, while the classic model fine-tuning on inherent parameters seems to be overlooked. It is believed that fine-tuning the parame
Externí odkaz:
http://arxiv.org/abs/2409.16718
Autor:
Shen, Xianjin, Liu, Hong-Li, Ren, Zhiyuan, Tej, Anandmayee, Li, Di, Liu, Hauyu Baobab, Fuller, Gary A., Xie, Jinjin, Jiao, Sihan, Yang, Aiyuan, Koch, Patrick M., Xu, Fengwei, Sanhueza, Patricio, Diep, Pham N., Peretto, Nicolas, Yadav, Ram K., Kramer, Busaba H., Sugiyama, Koichiro, Rawlings, Mark, Lee, Chang Won, Tatematsu, Ken'ichi, Harsono, Daniel, Eden, David, Kwon, Woojin, Tsai, Chao-Wei, White, Glenn, Kim, Kee-Tae, Liu, Tie, Wang, Ke, Zhang, Siju, Jiao, Wenyu, Yang, Dongting, Swagat, Das R., Wu, Jingwen, Wang, Chen
Filaments are believed to play a key role in high-mass star formation. We present a systematic study of the filaments and their hosting clumps in the G35 molecular complex using JCMT SCUBA-2 850 $\micron$ continuum data. We identified five clouds in
Externí odkaz:
http://arxiv.org/abs/2409.05492
This paper addresses user-specific dialogs. In contrast to previous research on personalized dialogue focused on achieving virtual user dialogue as defined by persona descriptions, user-specific dialogue aims to reproduce real-user dialogue beyond pe
Externí odkaz:
http://arxiv.org/abs/2409.00887
Autor:
Sugiyama, Kosuke, Uchida, Masato
While precise data observation is essential for the learning processes of predictive models, it can be challenging owing to factors such as insufficient observation accuracy, high collection costs, and privacy constraints. In this paper, we examines
Externí odkaz:
http://arxiv.org/abs/2408.14788
In large-data applications, such as the inference process of diffusion models, it is desirable to design sampling algorithms with a high degree of parallelization. In this work, we study the adaptive complexity of sampling, which is the minimal numbe
Externí odkaz:
http://arxiv.org/abs/2408.13045
Autor:
Kikuchi, Hirotsugu, Nishikawa, Hiroya, Matsukizono, Hiroyuki, Iino, Shunpei, Sugiyama, Takeharu, Ishioka, Toshio, Okumura, Yasushi
In our previous study, we demonstrated the existence of an unidentified ferroelectric smectic phase in the low-temperature region of the ferroelectric smectic A phase, where the layer spacing decreases with decreasing temperature. In the present stud
Externí odkaz:
http://arxiv.org/abs/2408.09520
Autor:
Delamarre, Amaury, Lombez, Laurent, Watanabe, Kentaroh, Sugiyama, Masakazu, Nakano, Yoshiaki, Guillemoles, Jean-Francois
Publikováno v:
IEEE Journal of Photovoltaics, 2016, 6 (2), pp.528-531
A recently suggested reciprocity relation states that the current transport efficiency from the junction to the cell terminal can be determined by differentiating luminescence images with respect to the terminal voltage. The validity of this relation
Externí odkaz:
http://arxiv.org/abs/2408.02463
Semi-supervised multi-label learning (SSMLL) is a powerful framework for leveraging unlabeled data to reduce the expensive cost of collecting precise multi-label annotations. Unlike semi-supervised learning, one cannot select the most probable label
Externí odkaz:
http://arxiv.org/abs/2407.18624
In this study, we investigate the decoherence of a spatially superposed electrically neutral spin-$\frac12$ particle in the presence of a relativistic quantum electromagnetic field in Minkowski spacetime. We demonstrate that decoherence due to the sp
Externí odkaz:
http://arxiv.org/abs/2407.14581
Autor:
LLM-jp, Aizawa, Akiko, Aramaki, Eiji, Chen, Bowen, Cheng, Fei, Deguchi, Hiroyuki, Enomoto, Rintaro, Fujii, Kazuki, Fukumoto, Kensuke, Fukushima, Takuya, Han, Namgi, Harada, Yuto, Hashimoto, Chikara, Hiraoka, Tatsuya, Hisada, Shohei, Hosokawa, Sosuke, Jie, Lu, Kamata, Keisuke, Kanazawa, Teruhito, Kanezashi, Hiroki, Kataoka, Hiroshi, Katsumata, Satoru, Kawahara, Daisuke, Kawano, Seiya, Keyaki, Atsushi, Kiryu, Keisuke, Kiyomaru, Hirokazu, Kodama, Takashi, Kubo, Takahiro, Kuga, Yohei, Kumon, Ryoma, Kurita, Shuhei, Kurohashi, Sadao, Li, Conglong, Maekawa, Taiki, Matsuda, Hiroshi, Miyao, Yusuke, Mizuki, Kentaro, Mizuki, Sakae, Murawaki, Yugo, Nakamura, Ryo, Nakamura, Taishi, Nakayama, Kouta, Nakazato, Tomoka, Niitsuma, Takuro, Nishitoba, Jiro, Oda, Yusuke, Ogawa, Hayato, Okamoto, Takumi, Okazaki, Naoaki, Oseki, Yohei, Ozaki, Shintaro, Ryu, Koki, Rzepka, Rafal, Sakaguchi, Keisuke, Sasaki, Shota, Sekine, Satoshi, Suda, Kohei, Sugawara, Saku, Sugiura, Issa, Sugiyama, Hiroaki, Suzuki, Hisami, Suzuki, Jun, Suzumura, Toyotaro, Tachibana, Kensuke, Takagi, Yu, Takami, Kyosuke, Takeda, Koichi, Takeshita, Masashi, Tanaka, Masahiro, Taura, Kenjiro, Tolmachev, Arseny, Ueda, Nobuhiro, Wan, Zhen, Yada, Shuntaro, Yahata, Sakiko, Yamamoto, Yuya, Yamauchi, Yusuke, Yanaka, Hitomi, Yokota, Rio, Yoshino, Koichiro
This paper introduces LLM-jp, a cross-organizational project for the research and development of Japanese large language models (LLMs). LLM-jp aims to develop open-source and strong Japanese LLMs, and as of this writing, more than 1,500 participants
Externí odkaz:
http://arxiv.org/abs/2407.03963