Accurate and efficient asset matching is crucial for remote management and diversified investment entities. To address the challenges of low efficiency, high labor costs, and false matching, this paper proposes a real...
详细信息
Electroencephalography (EEG) data is invaluable for investigating EEG physiology, neuroscience and clinical medicine. However, EEG signals are characterized by temporal nature, high dimensionality, and noise, which ma...
详细信息
Introducing LexiFuse+: a semi-supervised model merging lexicon features, BERT transfer learning, and one-class classifiers to detect anomalous content in short texts. It tackles challenges of informal text and imbalan...
详细信息
UHI is the process whereby human activity and urban infrastructure cause urban regions to suffer from warmer temperatures than rural ones. This study explores the use of machine and deep learning models for temperatur...
详细信息
It has recently been discovered that using a pretrained vision-language model (VLM), e.g., CLIP, to align a whole query image with several finer text descriptions generated by a large language model can significantly ...
详细信息
It has recently been discovered that using a pretrained vision-language model (VLM), e.g., CLIP, to align a whole query image with several finer text descriptions generated by a large language model can significantly enhance zero-shot performance. However, in this paper, we empirically find that the finer descriptions tend to align more effectively with local areas of the query image rather than the whole image, and then we theoretically validate this finding. Thus, we present a method called weighted visual-text cross alignment (WCA). This method begins with a localized visual prompting technique, designed to identify local visual areas within the query image. The local visual areas are then cross-aligned with the finer descriptions by creating a similarity matrix using the pre-trained VLM. To determine how well a query image aligns with each category, we develop a score function based on the weighted similarities in this matrix. Extensive experiments demonstrate that our method significantly improves zero-shot performance across various datasets, achieving results that are even comparable to few-shot learning methods. The code is available at ***/tmlr-group/WCA. Copyright 2024 by the author(s)
Distributed generators (DGs) are considered as significant components to modern micro grids because they can provide instant and renewable electric power to consumers without using transmission networks. However, the ...
详细信息
Driver behaviour modelling is a critical field that addresses complex and dynamic driving behaviours on roads with the goal of enhancing road safety, reducing air pollution, and improving vehi...
详细信息
With the booming of cyber attacks and cyber criminals against cyber-physical systems(CPSs),detecting these attacks remains *** might be the worst of times,but it might be the best of times because of opportunities bro...
详细信息
With the booming of cyber attacks and cyber criminals against cyber-physical systems(CPSs),detecting these attacks remains *** might be the worst of times,but it might be the best of times because of opportunities brought by machine learning(ML),in particular deep learning(DL).In general,DL delivers superior performance to ML because of its layered setting and its effective algorithm for extract useful information from training *** models are adopted quickly to cyber attacks against CPS *** this survey,a holistic view of recently proposed DL solutions is provided to cyber attack detection in the CPS context.A six-step DL driven methodology is provided to summarize and analyze the surveyed literature for applying DL methods to detect cyber attacks against CPS *** methodology includes CPS scenario analysis,cyber attack identification,ML problem formulation,DL model customization,data acquisition for training,and performance *** reviewed works indicate great potential to detect cyber attacks against CPS through DL ***,excellent performance is achieved partly because of several highquality datasets that are readily available for public ***,challenges,opportunities,and research trends are pointed out for future research.
The explosive growth of online social networks (OSNs) has led to the emergence of a variety of communities, where users can share their interests, skills, and knowledge. In this context, identifying the most influenti...
详细信息
This paper aims to investigate the up-to-date functions and features offered in PHR systems, that might affect PHRs adoption, engagement, and usability. Article search was conducted in the Scopus, Medline, and Web of ...
详细信息
暂无评论