Zobrazeno 1 - 10
of 1 955
pro vyhledávání: '"LIU Ziyu"'
Autor:
Liu, Ziyu, Zang, Yuhang, Dong, Xiaoyi, Zhang, Pan, Cao, Yuhang, Duan, Haodong, He, Conghui, Xiong, Yuanjun, Lin, Dahua, Wang, Jiaqi
Visual preference alignment involves training Large Vision-Language Models (LVLMs) to predict human preferences between visual inputs. This is typically achieved by using labeled datasets of chosen/rejected pairs and employing optimization algorithms
Externí odkaz:
http://arxiv.org/abs/2410.17637
We study the large deviation principle (LDP) for locally damped nonlinear wave equations perturbed by a bounded noise. When the noise is sufficiently non-degenerate, we establish the LDP for empirical distributions with lower bound of a local type. T
Externí odkaz:
http://arxiv.org/abs/2409.11717
We establish a new criterion for exponential mixing of random dynamical systems. Our criterion is applicable to a wide range of systems, including in particular dispersive equations. Its verification is in nature related to several topics, i.e., asym
Externí odkaz:
http://arxiv.org/abs/2407.15058
ESP-MedSAM: Efficient Self-Prompting SAM for Universal Domain-Generalized Medical Image Segmentation
Autor:
Xu, Qing, Li, Jiaxuan, He, Xiangjian, Liu, Ziyu, Chen, Zhen, Duan, Wenting, Li, Chenxin, He, Maggie M., Tesema, Fiseha B., Cheah, Wooi P., Wang, Yi, Qu, Rong, Garibaldi, Jonathan M.
The universality of deep neural networks across different modalities and their generalization capabilities to unseen domains play an essential role in medical image segmentation. The recent Segment Anything Model (SAM) has demonstrated its potential
Externí odkaz:
http://arxiv.org/abs/2407.14153
Self-supervised contrastive learning has become a key technique in deep learning, particularly in time series analysis, due to its ability to learn meaningful representations without explicit supervision. Augmentation is a critical component in contr
Externí odkaz:
http://arxiv.org/abs/2407.09336
Autor:
Ma, Yubo, Zang, Yuhang, Chen, Liangyu, Chen, Meiqi, Jiao, Yizhu, Li, Xinze, Lu, Xinyuan, Liu, Ziyu, Ma, Yan, Dong, Xiaoyi, Zhang, Pan, Pan, Liangming, Jiang, Yu-Gang, Wang, Jiaqi, Cao, Yixin, Sun, Aixin
Understanding documents with rich layouts and multi-modal components is a long-standing and practical task. Recent Large Vision-Language Models (LVLMs) have made remarkable strides in various tasks, particularly in single-page document understanding
Externí odkaz:
http://arxiv.org/abs/2407.01523
Autor:
Liu, Ziyu, Chu, Tao, Zang, Yuhang, Wei, Xilin, Dong, Xiaoyi, Zhang, Pan, Liang, Zijian, Xiong, Yuanjun, Qiao, Yu, Lin, Dahua, Wang, Jiaqi
Generating natural and meaningful responses to communicate with multi-modal human inputs is a fundamental capability of Large Vision-Language Models(LVLMs). While current open-source LVLMs demonstrate promising performance in simplified scenarios suc
Externí odkaz:
http://arxiv.org/abs/2406.11833
Autor:
Wang, Jiaqi, Zang, Yuhang, Zhang, Pan, Chu, Tao, Cao, Yuhang, Sun, Zeyi, Liu, Ziyu, Dong, Xiaoyi, Wu, Tong, Lin, Dahua, Chen, Zeming, Wang, Zhi, Meng, Lingchen, Yao, Wenhao, Yang, Jianwei, Wu, Sihong, Chen, Zhineng, Wu, Zuxuan, Jiang, Yu-Gang, Wu, Peixi, Chai, Bosong, Nie, Xuan, Yan, Longquan, Wang, Zeyu, Zhou, Qifan, Wang, Boning, Huang, Jiaqi, Xu, Zunnan, Li, Xiu, Yuan, Kehong, Zu, Yanyan, Ha, Jiayao, Gao, Qiong, Jiao, Licheng
Detecting objects in real-world scenes is a complex task due to various challenges, including the vast range of object categories, and potential encounters with previously unknown or unseen objects. The challenges necessitate the development of publi
Externí odkaz:
http://arxiv.org/abs/2406.11739
Autor:
Jia, Junlong, Hu, Ying, Weng, Xi, Shi, Yiming, Li, Miao, Zhang, Xingjian, Zhou, Baichuan, Liu, Ziyu, Luo, Jie, Huang, Lei, Wu, Ji
We present TinyLLaVA Factory, an open-source modular codebase for small-scale large multimodal models (LMMs) with a focus on simplicity of code implementations, extensibility of new features, and reproducibility of training results. Following the des
Externí odkaz:
http://arxiv.org/abs/2405.11788
Publikováno v:
DAC2024
Trusted Execution Environments (TEEs) have become a promising solution to secure DNN models on edge devices. However, the existing solutions either provide inadequate protection or introduce large performance overhead. Taking both security and perfor
Externí odkaz:
http://arxiv.org/abs/2405.03974