检索结果-内蒙古大学图书馆

arXiv 2021年

作者： Lyu, Boer Chen, Lu Zhu, Su Yu, Kai State Key Laboratory of Media Convergence Production Technology and Systems MoE Key Lab of Artificial Intelligence AI Institute Shanghai Jiao Tong University X-LANCE Lab Department of Computer Science and Engineering Shanghai Jiao Tong University Shanghai China

Chinese short text matching is a fundamental task in natural language processing. Existing approaches usually take Chinese characters or words as input tokens. They have two limitations: 1) Some Chinese words are polysemous, and semantic information is not fully utilized. 2) Some models suffer potential issues caused by word segmentation. Here we introduce HowNet as an external knowledge base and propose a Linguistic knowledge Enhanced graph Transformer (LET) to deal with word ambiguity. Additionally, we adopt the word lattice graph as input to maintain multi-granularity information. Our model is also complementary to pre-trained language models. Experimental results on two Chinese datasets show that our models outperform various typical text matching approaches. Ablation study also indicates that both semantic information and multi-granularity information are important for text matching modeling. Copyright © 2021, The Authors. All rights reserved.

关键词： Semantics

来源：评论

学校读者我要写书评

暂无评论

Efficient Image Super-Resolution with Feature Interaction Weighted Hybrid Network

arXiv

引用

arXiv 2022年

作者： Li, Wenjie Li, Juncheng Gao, Guangwei Deng, Weihong Yang, Jian Qi, Guo-Jun Lin, Chia-Wen Pattern Recognition and Intelligent System Laboratory School of Artificial Intelligence Beijing University of Posts and Telecommunications Beijing100080 China Intelligent Visual Information Perception Laboratory Institute of Advanced Technology Nanjing University of Posts and Telecommunications Nanjing210046 China Key Laboratory of Artificial Intelligence Ministry of Education Shanghai200240 China Provincial Key Laboratory for Computer Information Processing Technology Soochow University Suzhou215006 China School of Communication and Information Engineering Shanghai University Shanghai200444 China School of Computer Science and Technology Nanjing University of Science and Technology Nanjing210094 China Research Center for Industries of the Future the School of Engineering Westlake University Hangzhou310024 China OPPO Research SeattleWA98101 United States Department of Electrical Engineering National Tsing Hua University Hsinchu30013 Taiwan

Lightweight image super-resolution aims to reconstruct high-resolution images from low-resolution images using low computational costs. However, existing methods result in the loss of middle-layer features due to activation functions. To minimize the impact of intermediate feature loss on reconstruction quality, we propose a Feature Interaction Weighted Hybrid Network (FIWHN), which comprises a series of Wide-residual Distillation Interaction Block (WDIB) as the backbone. Every third WDIB forms a Feature Shuffle Weighted Group (FSWG) by applying mutual information shuffle and fusion. Moreover, to mitigate the negative effects of intermediate feature loss, we introduce Wide Residual Weighting units within WDIB. These units effectively fuse features of varying levels of detail through a Wide-residual Distillation Connection (WRDC) and a Self-Calibrating Fusion (SCF). To compensate for global feature deficiencies, we incorporate a Transformer and explore a novel architecture to combine CNN and Transformer. We show that our FIWHN achieves a favorable balance between performance and efficiency through extensive experiments on low-level and high-level tasks. Codes will be available at https://***/IVIPLab/FIWHN. Copyright © 2022, The Authors. All rights reserved.

关键词： Image reconstruction

来源：评论

学校读者我要写书评

暂无评论

Improving Referring Expression Comprehension by Suppressing Expression-unrelated Proposals

Improving Referring Expression Comprehension by Suppressing ...

引用

Electronic Information Engineering and computer technology (EIECT), International Conference on

作者： Qun Wang Feng Zhu Xiang Li Ruida Ye Key Laboratory of Opto-Electronic Information Processing Chinese Academy of Sciences Shenyang China Shenyang Institute of Automation Chinese Academy of Sciences Shenyang China Institutes for Robotics and Intelligent Manufacturing Chinese Academy of Sciences Shenyang China University of Chinese Academy of Sciences Beijing China Faculty of Computer Science and Engineering Northeastern University Shenyang China Department of Aerospace Engineering and Technology Space Engineering University Beijing China

The prevailing referring expression comprehension (REC) method is based on two stages: 1) detecting multiple candidate proposals using an object detector and 2) aligning proposals with text descriptions. Existing two-stage models have achieved significant improvements with excellent alignment modules. However, their alignment modules treat each proposal equally without filtering expression-unrelated proposals. These expression-unrelated proposals may interfere with the alignment modules’ ability to accurately reason about the target proposal. These proposals also increase the difficulty of training alignment modules. To solve this problem, we propose a new framework, that can suppress expression-unrelated proposals for alignment modules. Specifically, we first design a suppression module, which calculates the relevance scores between the proposals and the expression by using the semantics in the expression. Then, these scores are integrated into three basic alignment modules, including subject, location, and relationship modules, to guide each module to focus on the expression-related proposals. We conduct extensive experiments on three benchmark datasets, and the results show the effectiveness of our suppression module and performance improvement.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Automatic Modulation Classification of Frequency-Hopping Signals Using High-Dimensional Phase Diagrams

Automatic Modulation Classification of Frequency-Hopping Sig...

引用

IEEE International Symposium on Broadband Multimedia systems and Broadcasting (BMSB)

作者： Qibo Chen Kun Yan Hsiao-Chun Wu Shih Yu Chang Xiao Yan Yiyan Wu Haonan Chang Guangxi Key Laboratory of Wireless Wideband Communication and Signal Processing Guilin University of Electronic Technology Guilin China Department of Information and Telecommunication Guilin University of Electronic Technology Guilin Guangxi China School of Electrical Engineering and Computer Science Louisiana State University Baton Rouge LA USA Department of Applied Data Science San Jose State University San Jose CA USA School of Aeronautics and Astronautics University of Electronic Science and Technology of China Chengdu P. R. China Communications Research Centre Canada Ottawa Canada Haonan Chang is with Xi'an University Xi'an P. R. China

Automatic modulation recognition for frequency-hopping (FH) signals remains very challenging to researchers due to the signals' time-varying spectral characteristics. In this work, a novel robust automatic modulation recognition scheme is investigated for FH signals using the phase-space topological features represented by the embedded phase diagrams. As such embedded phase diagrams are often high-dimensional, it is necessary to formulate the phase-space features as tensors. In the training process, the phase-space tensor features will be utilized to establish the regression models as linear encoders for the individual modulations. The aforementioned linear encoders are constructed using the support vector machine (SVM); the phase-space feature-tensors of the training signals of all modulations will be projected by their corresponding regression models (or linearly encoded) to produce the representative code-vectors, respectively. In the test stage, the phase-space feature-tensor produced from a test signal will be projected by each individual trained regression model (or linearly encoded) to generate the respective code-vectors. Then, the code-vectors resulting from the test stage will be compared with the representative code-vectors to find which modulation will lead to the smallest Euclidean distance in between and such a modulation will be picked as the modulation type of the test signal. Monte Carlo simulation results have demonstrated that the average recognition accuracy of our proposed new approach is more than 90% when the signal-to-noise ratio is no less than 0 dB for additive white Gaussian noise.

关键词：

来源：评论

学校读者我要写书评

暂无评论

In vivo high-resolution multimodal nonlinear optical microscopy of spinal cord in mice

In vivo high-resolution multimodal nonlinear optical microsc...

引用

Advanced Chemical Microscopy for Life science and Translational Medicine 2020

作者： Wu, Wanjie Qin, Zhongya Wu, Junqiang Chen, Congping Chen, Weitao Liu, Kai Qu, Jianan Y. Biophotonics Research Laboratory Department of Electronic and Computer Engineering Hong Kong University of Science and Technology Clear Water Bay Kowloon Hong Kong Division of Life Science State Key Laboratory of Molecular Neuroscience Hong Kong University of Science and Technology Clear Water Bay Kowloon Hong Kong Center of Systems Biology and Human Health Hong Kong University of Science and Technology Clear Water Bay Kowloon Hong Kong

ISBN: (纸本)9781510632677

Chronic in vivo optical imaging of the spinal cord is an effective way to study the biological processes during and after spinal cord injury (SCI) in mouse models. It normally relies on an implanted spinal chamber to provide continuous optical access to the spinal cord. However, the chronic window consists of multiple layers of transparent materials with various optical properties and irregular thickness, which induce large optical aberration. Therefore, the image quality of multiphoton microscopy as well as the precision of femtosecond laser axotomy were dramatically degraded. In this work, we developed an adaptive optics (AO) microscope system integrating stimulated Raman scattering (SRS) and twophoton excited fluorescence (TPEF). Using our system, the aberrations induced by the spinal cord window were measured and compensated accordingly, enabling both high-resolution imaging and precise laser axotomy of the mouse spinal cord. © COPYRIGHT SPIE. Downloading of the abstract is permitted for personal use only.

关键词： Adaptive optics

来源：评论

学校读者我要写书评

暂无评论

Large Models Defining a New Era in Intelligence Testing: A Short Discussion and Exploration

IEEE Transactions on Intelligent Vehicles

引用

IEEE Transactions on intelligent Vehicles 2024年第8期9卷 5247-5249页

作者： Rui Zhou Jingwei Ge Jing Yang Peijun Ye Yuchen Li Levente Kovács Macau University of Science and Technology Macau China Waytous Inc. Shenzhen China Department of Automation BNRist Tsinghua University Beijing China State Key Laboratory of Multimodal Artificial Intelligence Systems Institute of Automation Chinese Academy of Sciences Beijing China School of Artificial Intelligence University of Chinese Academy of Sciences Beijing China Department of Informatics Technische Universität München München Germany Robotics College of Óbuda University Budapest Hungary John von Neumann Faculty of Informatics Obuda University Budapest Hungary

This perspective presents a short yet comprehensive discussion on the rapid advancements in the field of intelligent system testing enhancing by large models. We delve into the pivotal aspects of testing, elucidating the roles of the Testing Operating System (TOS), Testing Foundational Intelligence (TFI), and Testing Scenario Engineering (TSE), and also the challenges they face from a testing perspective.

关键词： Testing Germanium Operating systems Data models Large language models Hardware Vehicle dynamics Biological system modeling Accuracy Scenario generation

来源：评论

学校读者我要写书评

暂无评论

SMedBERT: A knowledge-enhanced pre-trained language model with structured semantics for medical text mining

arXiv

引用

arXiv 2021年

作者： Zhang, Taolin Cai, Zerui Wang, Chengyu Qiu, Minghui Yang, Bite He, Xiaofeng School of Software Engineering East China Normal University Alibaba Group Shanghai Key Laboratory of Trsustworthy Computing School of Computer Science and Technology East China Normal University Dxy Shanghai Research Institute for Intelligent Autonomous Systems

Recently, the performance of Pre-trained Language Models (PLMs) has been significantly improved by injecting knowledge facts to enhance their abilities of language understanding. For medical domains, the background knowledge sources are especially useful, due to the massive medical terms and their complicated relations are difficult to understand in text. In this work, we introduce SMedBERT, a medical PLM trained on large-scale medical corpora, incorporating deep structured semantics knowledge from neighbours of linked-entity. In SMedBERT, the mention-neighbour hybrid attention is proposed to learn heterogeneousentity information, which infuses the semantic representations of entity types into the homogeneous neighbouring entity structure. Apart from knowledge integration as external features, we propose to employ the neighbors of linked-entities in the knowledge graph as additional global contexts of text mentions, allowing themto communicate via shared neighbors, thus enrich their semantic representations. Experiments demonstrate that SMedBERT significantly outperforms strong baselines in various knowledge-intensive Chinese medical tasks. It also improves the performance of other tasks such as question answering, questionmatching and natural language inference. Copyright © 2021, The Authors. All rights reserved.

关键词： Knowledge graph

来源：评论

学校读者我要写书评

暂无评论

Error Correction of Tibetan Verbs Based on Deep Learning

Error Correction of Tibetan Verbs Based on Deep Learning

引用

Artificial Intelligence and computer Engineering (ICAICE), International Conference on

作者： HUA Guo‒ cai– rang Secha Jia BAN Ma– bao CAI Rang– jia School of computer science Qinghai Normal University China Qinghai Key Laboratory of Tibetan information processing and machine translation China Qinghai Tibetan information processing engineering technology research center China State Key Laboratory of Tibetan intelligent information processing and application China

Verbs are the core of the semantic structure represented by Tibetan sentences, and automatic error correction of Tibetan verbs is one of the important research topics in Tibetan language processing. In this paper, we propose a Bi-LSTM neural network model for Tibetan verb error correction by analyzing the usage rules of verbs in Tibetan, summarizing the grammatical, semantic and spelling features of verbs, and proposes a Bi-LSTM neural network model for Tibetan verb error correction based on these features, which not only extracts various features of verbs, but also capture the contextual information of verbs through Bi-LSTM neural network. The method proposed in this work solves the drawbacks of the traditional methods of low generalization and inability to obtain long-distance contextual implicit information. The experimental results show that the accuracy, recall and F1 values of the proposed method on the test set reach 97.3%,95.7% and 96.9%, respectively, indicating the effectiveness of the proposed method on the task of automatic error correction for Tibetan verbs.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Towards Sample-specific Backdoor Attack with Clean Labels via Attribute Trigger

arXiv

引用

arXiv 2023年

作者： Li, Yiming Zhu, Mingyan Guo, Junfeng Wei, Tao Xia, Shu-Tao Qin, Zhan Hangzhou311215 China School of Cyber Science and Technology Zhejiang University Hangzhou311200 China Tsinghua Shenzhen International Graduate School Tsinghua University Shenzhen518055 China Department of Computer Science University of Maryland College ParkMD20742 United States Ant Group Hangzhou310023 China The Research Center of Artificial Intelligence Peng Cheng Laboratory Shenzhen518000 China The State Key Laboratory of Blockchain and Data Security Zhejiang University Hangzhou311200 China

Currently, sample-specific backdoor attacks (SSBAs) are the most advanced and malicious methods since they can easily circumvent most of the current backdoor defenses. In this paper, we reveal that SSBAs are not sufficiently stealthy due to their poisoned-label nature, where users can discover anomalies if they check the image-label relationship. In particular, we demonstrate that it is ineffective to directly generalize existing SSBAs to their clean-label variants by poisoning samples solely from the target class. We reveal that it is primarily due to two reasons, including (1) the 'antagonistic effects' of ground-truth features and (2) the learning difficulty of sample-specific features. Accordingly, trigger-related features of existing SSBAs cannot be effectively learned under the clean-label setting due to their mild trigger intensity required for ensuring stealthiness. We argue that the intensity constraint of existing SSBAs is mostly because their trigger patterns are 'content-irrelevant' and therefore act as 'noises' for both humans and DNNs. Motivated by this understanding, we propose to exploit content-relevant features, a.k.a. (human-relied) attributes, as the trigger patterns to design clean-label SSBAs. This new attack paradigm is dubbed backdoor attack with attribute trigger (BAAT). Extensive experiments are conducted on benchmark datasets, which verify the effectiveness of our BAAT and its resistance to existing defenses. © 2023, CC BY.

关键词：

来源：评论

学校读者我要写书评

暂无评论

An open-source project for real-time image semantic segmentation

引用

science China(Information sciences) 2019年第12期62卷 246-247页

作者： Quan ZHOU Yu WANG Jia LIU Xin JIN Longin Jan LATECKI National Engineering Research Center of Communication and Network Technology Nanjing University of Posts and Telecommunications Key Laboratory of Ministry of Education for Broadband Wireless Communication and Sensor Network Technology Nanjing University of Posts and Telecommunications Beijing Electronic Science and Technology Institute State Key Laboratory of Virtual Reality Technology and Systems Beihang University Department of Computer and Information Sciences Temple University

Recent years have witnessed great progress of deep convolutional neural networks (DCNNs) for solving scene understanding tasks [1–3]. These advances prefer to construct deeper and larger net-

关键词： image An open-source project for real-time image semantic segmentation

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：