Zobrazeno 1 - 10
of 1 383
pro vyhledávání: '"Wang, Zhixiang"'
Controllable magnetization in atomically thin two-dimensional magnets is highly desirable for developing spintronics. For FeSe monolayer, its magnetic ground state is not yet fully understood, and the potential in constructing high-speed and advanced
Externí odkaz:
http://arxiv.org/abs/2409.07910
With the help of in-context learning (ICL), large language models (LLMs) have achieved impressive performance across various tasks. However, the function of descriptive instructions during ICL remains under-explored. In this work, we propose an ensem
Externí odkaz:
http://arxiv.org/abs/2408.08780
In-context learning (ICL) greatly improves the performance of large language models (LLMs) on various down-stream tasks, where the improvement highly depends on the quality of demonstrations. In this work, we introduce syntactic knowledge to select b
Externí odkaz:
http://arxiv.org/abs/2408.04872
Autor:
Wang, Zhixiang, Li, Baiang, Wang, Jian, Liu, Yu-Lun, Gu, Jinwei, Chuang, Yung-Yu, Satoh, Shin'ichi
This paper introduces an innovative approach for image matting that redefines the traditional regression-based task as a generative modeling challenge. Our method harnesses the capabilities of latent diffusion models, enriched with extensive pre-trai
Externí odkaz:
http://arxiv.org/abs/2407.21017
Autor:
Yeh, Chang-Han, Lin, Chin-Yang, Wang, Zhixiang, Hsiao, Chi-Wei, Chen, Ting-Hsuan, Shiu, Hau-Shiang, Liu, Yu-Lun
This paper introduces a method for zero-shot video restoration using pre-trained image restoration diffusion models. Traditional video restoration methods often need retraining for different settings and struggle with limited generalization across va
Externí odkaz:
http://arxiv.org/abs/2407.01519
In-context learning (ICL) is the trending prompting strategy in the era of large language models (LLMs), where a few examples are demonstrated to evoke LLMs' power for a given task. How to select informative examples remains an open issue. Previous w
Externí odkaz:
http://arxiv.org/abs/2403.19285
Event cameras are bio-inspired vision sensors that asynchronously measure per-pixel brightness changes. The high temporal resolution and asynchronicity of event cameras offer great potential for estimating the robot motion state. Recent works have ad
Externí odkaz:
http://arxiv.org/abs/2402.16398
Coreset selection seeks to choose a subset of crucial training samples for efficient learning. It has gained traction in deep learning, particularly with the surge in training dataset sizes. Sample selection hinges on two main aspects: a sample's rep
Externí odkaz:
http://arxiv.org/abs/2401.16193
We present HumanNeRF-SE, a simple yet effective method that synthesizes diverse novel pose images with simple input. Previous HumanNeRF works require a large number of optimizable parameters to fit the human images. Instead, we reload these approache
Externí odkaz:
http://arxiv.org/abs/2312.02232
Clinically, automated polyp segmentation techniques have the potential to significantly improve the efficiency and accuracy of medical diagnosis, thereby reducing the risk of colorectal cancer in patients. Unfortunately, existing methods suffer from
Externí odkaz:
http://arxiv.org/abs/2309.10219