Zobrazeno 1 - 10
of 162
pro vyhledávání: '"Junming An"'
This paper advances a novel architectural schema anchored upon the Transformer paradigm and innovatively amalgamates the K-means categorization algorithm to augment the contextual apprehension capabilities of the schema. The transformer model perform
Externí odkaz:
http://arxiv.org/abs/2408.04216
Mix-up is a key technique for consistency regularization-based semi-supervised learning methods, generating strong-perturbed samples for strong-weak pseudo-supervision. Existing mix-up operations are performed either randomly or with predefined rules
Externí odkaz:
http://arxiv.org/abs/2407.21586
This paper studies the active flux (AF) methods for two-dimensional hyperbolic conservation laws, focusing on the flux vector splitting (FVS) for the point value update and bound-preserving (BP) limitings, which is an extension of our previous work [
Externí odkaz:
http://arxiv.org/abs/2407.13380
Autor:
Duan, Haodong, Yang, Junming, Qiao, Yuxuan, Fang, Xinyu, Chen, Lin, Liu, Yuan, Dong, Xiaoyi, Zang, Yuhang, Zhang, Pan, Wang, Jiaqi, Lin, Dahua, Chen, Kai
We present VLMEvalKit: an open-source toolkit for evaluating large multi-modality models based on PyTorch. The toolkit aims to provide a user-friendly and comprehensive framework for researchers and developers to evaluate existing multi-modality mode
Externí odkaz:
http://arxiv.org/abs/2407.11691
The existing barely-supervised medical image segmentation (BSS) methods, adopting a registration-segmentation paradigm, aim to learn from data with very few annotations to mitigate the extreme label scarcity problem. However, this paradigm poses a ch
Externí odkaz:
http://arxiv.org/abs/2407.05248
Few-shot keyword spotting (KWS) aims to detect unknown keywords with limited training samples. A commonly used approach is the pre-training and fine-tuning framework. While effective in clean conditions, this approach struggles with mixed keyword spo
Externí odkaz:
http://arxiv.org/abs/2407.06078
Autor:
Qiao, Yuxuan, Duan, Haodong, Fang, Xinyu, Yang, Junming, Chen, Lin, Zhang, Songyang, Wang, Jiaqi, Lin, Dahua, Chen, Kai
Vision Language Models (VLMs) demonstrate remarkable proficiency in addressing a wide array of visual questions, which requires strong perception and reasoning faculties. Assessing these two competencies independently is crucial for model refinement,
Externí odkaz:
http://arxiv.org/abs/2406.14544
Text-to-image generation models often struggle with key element loss or semantic confusion in tasks involving Chinese classical poetry.Addressing this issue through fine-tuning models needs considerable training costs. Additionally, manual prompts fo
Externí odkaz:
http://arxiv.org/abs/2407.06196
Autor:
Zhang, Junming
We prove the representation given by a stable $\alpha_1$-cyclic parabolic $\mathrm{SO}_0(2,3)$-Higgs bundle through the non-Abelian Hodge correspondence is $\{\alpha_2\}$-almost dominated. This is a generalization of Filip's result on weight $3$ vari
Externí odkaz:
http://arxiv.org/abs/2406.08118
DR-RAG: Applying Dynamic Document Relevance to Retrieval-Augmented Generation for Question-Answering
Autor:
Hei, Zijian, Liu, Weiling, Ou, Wenjie, Qiao, Juyi, Jiao, Junming, Song, Guowen, Tian, Ting, Lin, Yi
Retrieval-Augmented Generation (RAG) has recently demonstrated the performance of Large Language Models (LLMs) in the knowledge-intensive tasks such as Question-Answering (QA). RAG expands the query context by incorporating external knowledge bases t
Externí odkaz:
http://arxiv.org/abs/2406.07348