Výsledky vyhledávání

Report

Re-boosting Self-Collaboration Parallel Prompt GAN for Unsupervised Image Restoration

Autor: Lin, Xin, Zhou, Yuyan, Yue, Jingtong, Ren, Chao, Chan, Kelvin C. K., Qi, Lu, Yang, Ming-Hsuan

Unsupervised restoration approaches based on generative adversarial networks (GANs) offer a promising solution without requiring paired datasets. Yet, these GAN-based approaches struggle to surpass the performance of conventional unsupervised GAN-bas

Externí odkaz: http://arxiv.org/abs/2408.09241

Zobrazit plný text záznamu

Report

MetaGPT: Merging Large Language Models Using Model Exclusive Task Arithmetic

Autor: Zhou, Yuyan, Song, Liang, Wang, Bingning, Chen, Weipeng

The advent of large language models (LLMs) like GPT-4 has catalyzed the exploration of multi-task learning (MTL), in which a single model demonstrates proficiency across diverse tasks. Task arithmetic has emerged as a cost-effective approach for MTL.

Externí odkaz: http://arxiv.org/abs/2406.11385

Zobrazit plný text záznamu

Report

Improving Generalization of Deep Neural Networks by Optimum Shifting

Autor: Zhou, Yuyan, Li, Ye, Feng, Lei, Huang, Sheng-Jun

Recent studies showed that the generalization of neural networks is correlated with the sharpness of the loss landscape, and flat minima suggests a better generalization ability than sharp minima. In this paper, we propose a novel method called \emph

Externí odkaz: http://arxiv.org/abs/2405.14111

Zobrazit plný text záznamu

Report

Improving Lens Flare Removal with General Purpose Pipeline and Multiple Light Sources Recovery

Autor: Zhou, Yuyan, Liang, Dong, Chen, Songcan, Huang, Sheng-Jun, Yang, Shuo, Li, Chongyi

When taking images against strong light sources, the resulting images often contain heterogeneous flare artifacts. These artifacts can importantly affect image visual quality and downstream computer vision tasks. While collecting real data pairs of f

Externí odkaz: http://arxiv.org/abs/2308.16460

Zobrazit plný text záznamu

Akademický článek

Ten Questions on Adjustable Water Volume of the Western Route of South-to-North Water Diversion Project

Autor: Hu Peng, Wang Hao, Zhao Yong, Ning Yuan, Jiang Yunzhong, Liu Huan, Zeng Qinghui, Yang Zefan, Zhou Yuyan, Dong Ningpeng, Yan Long, A Yinglan, Zhang Fengbo, Tang Jiaxuan, Wang Yulian, Wang Jianhua

Publikováno v: 中国工程科学, Vol 26, Iss 2, Pp 210-223 (2024)

The Western Route of the South-to-North Water Diversion Project is the final link to be constructed for the four-horizontal three-vertical framework of China’s national water network. The adjustable water volume in the water source areas is a focal

Externí odkaz: https://doaj.org/article/d28390b006194ce3a4de7abb0842890b