Zobrazeno 1 - 10
of 1 752
pro vyhledávání: '"FAN, Ting"'
An ideal length-extrapolatable Transformer language model can handle sequences longer than the training length without any fine-tuning. Such long-context utilization capability relies heavily on a flexible positional embedding design. Upon investigat
Externí odkaz:
http://arxiv.org/abs/2311.00684
Autor:
Fan, Ting-Yu
OBJECTIVE: Obesity is a chronic disease with high incidence worldwide, which promotes the risk of incidence of type 2 diabetes (T2D). Obesity-induced adipocyte expansion promotes local chronic inflammation in the adipose tissue which is considered a
Externí odkaz:
https://hdl.handle.net/2144/47451
In recent studies, linear recurrent neural networks (LRNNs) have achieved Transformer-level performance in natural language and long-range modeling, while offering rapid parallel training and constant inference cost. With the resurgence of interest i
Externí odkaz:
http://arxiv.org/abs/2309.07412
The use of positional embeddings in transformer language models is widely accepted. However, recent research has called into question the necessity of such embeddings. We further extend this inquiry by demonstrating that a randomly initialized and fr
Externí odkaz:
http://arxiv.org/abs/2305.13571
Unlike recurrent models, conventional wisdom has it that Transformers cannot perfectly model regular languages. Inspired by the notion of working memory, we propose a new Transformer variant named RegularGPT. With its novel combination of Weight-Shar
Externí odkaz:
http://arxiv.org/abs/2305.03796
Autor:
I-Hung Shao, Fan-Ting Liao, Chun-Bi Chang, Ying-Hsu Chang, Li-Jen Wang, Liang-Kang Huang, Hung-Cheng Kan, Po-Hung Lin, Kai-Jie Yu, Cheng-Keng Chuang, Chun-Te Wu, See-Tong Pang
Publikováno v:
Scientific Reports, Vol 14, Iss 1, Pp 1-9 (2024)
Abstract MRI-guided targeted biopsy (MRGB) was recommended as part of biopsy paradigm of prostate cancers by current guidelines. This study aimed to analyze the diagnostic efficacy of MRGB and systemic biopsy (SB), and to compare diagnostic capabilit
Externí odkaz:
https://doaj.org/article/7d3662bb55d64415bc8f1f4de7756047
Length extrapolation permits training a transformer language model on short sequences that preserves perplexities when tested on substantially longer sequences. A relative positional embedding design, ALiBi, has had the widest usage to date. We disse
Externí odkaz:
http://arxiv.org/abs/2212.10356
Autor:
Fan Ting, Zhang Zeyi
Publikováno v:
Frontiers in Bioengineering and Biotechnology, Vol 12 (2024)
PurposeThis study aimed to analyze the biomechanical characteristics of the lower limb in patients with patellofemoral pain (PFP) while walking under different sensory integration tasks and elucidate the relationship between these biomechanical chara
Externí odkaz:
https://doaj.org/article/d1b08dfd66714cbc93ee5ac5ff23241d
Publikováno v:
Symphonya, Iss 1 (2024)
This study intends to determine the elements for adopting corporate social responsibility (CSR) practices that would boost the organization's performance (OP) either directly or indirectly in Pakistan's oil and gas sectors. This research demonstrates
Externí odkaz:
https://doaj.org/article/87ae3e21fcc6445989ec8b0b8dd5f13d
While deep generative models have succeeded in image processing, natural language processing, and reinforcement learning, training that involves discrete random variables remains challenging due to the high variance of its gradient estimation process
Externí odkaz:
http://arxiv.org/abs/2206.07235