Qingyuan Hu

A comparative analysis of explainable artificial intelligence (XAI) models for predicting concrete elastic dynamic modulus

Haoyang Zheng, Yuxiang Huang, Kechang Wu, Bowen Wang et autres

The application of explainable artificial intelligence (XAI) in civil engineering has garnered increasing attention due to its ability to enhance transparency in machine learning (ML) models for material behavior prediction. However, the comparative performance and interpretability of various XAI models in predicting …

cn, jp, gb (code pays fourni par la source)

CSegNet: a hybrid transformer-CNN network for road crack image segmentation

Hao Dong, Yinlai Du, Qingyuan Hu, Mingzhu Zhou et autres

Detecting cracks from optical images plays a crucial role in road maintenance but its good realisation has many challenges. Road cracks exhibit significant diversity and complexity in terms of shape, size and texture and road images may contain various types of noise …

cn (code pays fourni par la source)

Graph Neural Network Model Under Multi-Modal Fusion

Tao Tao, Jibin Wang, Zhuo Chen, Qingyuan Hu et autres

In the real world, data usually exhibits multimodal characteristics with different sources and formats. Traditional unimodal processing methods often fail to fully utilize the correlation and complementary information between different modalities. In this paper, we integrate various modal data into a graph …

cn (code pays fourni par la source)

Energy-aware feature and classifier for behaviour recognition of laying hens in an aviary system

Xiao Bo Yang, Qingyuan Hu, Chunyu Wang

• Long-term monitoring of aviary birds’ behaviours is challenging. • A behaviour classification model for laying hens in aviary systems was proposed. • 1-s sliding window yielded higher accuracy than the 0.5-s window. • Denoising significantly improved data quality and model performance. …

cn (code pays fourni par la source)

Robustly packaged lithium niobate resonator-based temperature sensor

Liu Yang, Yongyong Zhuang, Yifan Zhang, Haochen Yan et autres

Abstract Whispering gallery mode (WGM) resonators with high optical quality factor (Q) are sensitive to environment changes and are used in various sensing applications. We have developed a high-precision temperature sensor using a compactly packaged lithium niobate WGM resonator. This sensor offers …

dk (code pays fourni par la source)

Data Catalogs with Artificial Intelligence and Active Metadata: A Case Study of China Mobile

Jibin Wang, Yanmei Guo, Yong Jiang, Jing Shang et autres

In the context of the rapid development of artificial intelligence, telecommunications operators have accumulated an increasing amount of data. China Mobile has developed a set of practical methods of enhanced data catalog in response to current issues such as difficulties in scheduling …

cn (code pays fourni par la source)

CSegNet：A Crack Segmentation Network Combining CNN and Transformer

Hao Dong, Yinlai Du, Dong Feng, Qingyuan Hu et autres

Abstract Detecting cracks from images plays a crucial role in road maintenance. Road cracks exhibit significant diversity and complexity in terms of shape, size, texture, and road images may contain various noises and interferences such as lighting variations, shadows, and different appearances …

cn (code pays fourni par la source)

PANDA: Preference Adaptation for Enhancing Domain-Specific Abilities of LLMs

An Liu, Zonghan Yang, Zhenhe Zhang, Qingyuan Hu et autres

While Large language models (LLMs) have demonstrated considerable capabilities across various natural language tasks, they often fall short of the performance achieved by domain-specific state-of-the-art models. One potential approach to enhance domain-specific capabilities of LLMs involves fine-tuning them using corresponding datasets. However, …

Attention-SA: Exploiting Model-Approximated Data Semantics for Adversarial Attack

Qian Li, Qingyuan Hu, Haoran Fan, Chenhao Lin et autres

Adversarial Defense of deep neural networks have gained significant attention and there have been active research efforts on model vulnerabilities for attacking such as gradient-based attack and pre-defined semantic manipulation. However, they often lack clear adversarial pattern connecting model extracted notion and …

cn (code pays fourni par la source)

PANDA: Preference Adaptation for Enhancing Domain-Specific Abilities of LLMs

An Liu, Zonghan Yang, Zhenhe Zhang, Qingyuan Hu et autres

While Large language models (LLMs) have demonstrated considerable capabilities across various natural language tasks, they often fall short of the performance achieved by domainspecific state-of-the-art models.One potential approach to enhance domain-specific capabilities of LLMs involves fine-tuning them using corresponding datasets.However, this method …

cn (code pays fourni par la source)

Complex Network Traffic Prediction Method Under Graph Neural Network

Qingyuan Hu, Tao Tao, Jibin Wang, Zhuo Chen et autres

cn (code pays fourni par la source)

Domain patterning in nonpolar cut PMN–PT by focused ion beam

E. A. Pashnina, D. S. Chezganov, Alla Slautina, Anton P. Turygin et autres

The formation of the ferroelectric domain structure as a result of irradiation by focused ion beam of [100]-cut 0.61Pb(Mg[Formula: see text]Nb[Formula: see text]O3–0.39PbTiO3 (PMN–PT) single crystals covered by surface artificial dielectric layer and with free surface was investigated. The dot irradiation resulted …

ru, cn (code pays fourni par la source)