Zobrazeno 1 - 10
of 107
pro vyhledávání: '"Xu, Shaoyuan"'
Autor:
Valentin, Simon, Fu, Jinmiao, Detommaso, Gianluca, Xu, Shaoyuan, Zappella, Giovanni, Wang, Bryan
Large language models (LLMs) can be prone to hallucinations - generating unreliable outputs that are unfaithful to their inputs, external facts or internally inconsistent. In this work, we address several challenges for post-hoc hallucination detecti
Externí odkaz:
http://arxiv.org/abs/2407.21424
This paper introduces \textbf{Q-tuning}, a novel approach for continual prompt tuning that enables the lifelong learning of a pre-trained language model. When learning a new task, Q-tuning trains a task-specific prompt by adding it to a prompt queue
Externí odkaz:
http://arxiv.org/abs/2404.14607
On shopping websites, product images of low quality negatively affect customer experience. Although there are plenty of work in detecting images with different defects, few efforts have been dedicated to correct those defects at scale. A major challe
Externí odkaz:
http://arxiv.org/abs/2309.05883
Semi-supervised learning (SSL) has become a crucial approach in deep learning as a way to address the challenge of limited labeled data. The success of deep neural networks heavily relies on the availability of large-scale high-quality labeled data.
Externí odkaz:
http://arxiv.org/abs/2309.05826
For image restoration, methods leveraging priors from generative models have been proposed and demonstrated a promising capacity to robustly restore photorealistic and high-quality results. However, these methods are susceptible to semantic ambiguity
Externí odkaz:
http://arxiv.org/abs/2203.03005
Autor:
Liu, Huidong, Xu, Shaoyuan, Fu, Jinmiao, Liu, Yang, Xie, Ning, Wang, Chien-Chih, Wang, Bryan, Sun, Yi
Modern Web systems such as social media and e-commerce contain rich contents expressed in images and text. Leveraging information from multi-modalities can improve the performance of machine learning tasks such as classification and recommendation. I
Externí odkaz:
http://arxiv.org/abs/2112.03562
The evolving algorithms for 2D facial landmark detection empower people to recognize faces, analyze facial expressions, etc. However, existing methods still encounter problems of unstable facial landmarks when applied to videos. Because previous rese
Externí odkaz:
http://arxiv.org/abs/2007.15269
Publikováno v:
Open Mathematics, Vol 21, Iss 1, Pp 133-181 (2023)
In this article, without requiring solidness of the underlying cone, a kind of new convergence for sequences in cone bb-metric spaces over Banach algebras and a new kind of completeness for such spaces, namely, wrtn-completeness, are introduced. Unde
Externí odkaz:
https://doaj.org/article/4cc495c3a3924b6e8a7072fe43d38943
Autor:
Xu, Shaoyuan1,2 (AUTHOR), Wang, Xiaoning1,2 (AUTHOR), Zhang, Ying1,2 (AUTHOR), Han, Yifan1,2 (AUTHOR), Zhang, Changjun1,2,3 (AUTHOR) li_jie2020@126.com
Publikováno v:
Gynecological Endocrinology. December 2023, Vol. 39 Issue 1, p1-5. 5p.
Autor:
Kang, Weiming, Tian, Jie, Lai, Yao, Xu, Shaoyuan, Gao, Chao, Hong, Weijie, Zhou, Yongxu, Pei, Lina, He, Chansheng
Publikováno v:
In Journal of Hydrology April 2022 607