检索结果-内蒙古大学图书馆

arXiv 2020年

作者： Kim, Nam Kyun Kim, Hong Kook School of Electrical Engineering and Computer Science AI Graduate School Gwangju Institute of Science and Technology 123 Cheomdangwagi-ro Gwangju61005 Korea Republic of

This report proposes a polyphonic sound event detection (SED) method for the DCASE 2020 Challenge Task 4. The proposed SED method is based on semi-supervised learning to deal with the different combination of training datasets such as weakly labeled dataset, unlabeled dataset, and strongly labeled synthetic dataset. Especially, the target label of each audio clip from weakly labeled or unlabeled dataset is first predicted by using the mean teacher model that is the DCASE 2020 baseline. The data with predicted labels are used for training the proposed SED model, which consists of CNNs with skip connections and self-attention mechanism, followed by RNNs. In order to compensate for the erroneous prediction of weakly labeled and unlabeled data, a semi-supervised loss function is employed for the proposed SED model. In this work, several versions of the proposed SED model are implemented and evaluated on the validation set according to the different parameter setting for the semi-supervised loss function, and then an ensemble model that combines five-fold validation models is finally selected as our final model. Copyright © 2020, The Authors. All rights reserved.

关键词： Recurrent neural networks

AI creativity and legal protection for AI-generated works in posthuman societal scenarios

学校读者我要写书评

暂无评论

Sustainable Futures 2025年 9卷

作者： Wei Li Ruiming Song Bing Zhang Kunpeng Yu School of Accounting Guangzhou College of Technology and Business Foshan 528138 PR China Interdepartmental Graduate Program in Marine Science University of California Santa Barbara Santa Barbara CA 93106 USA School of Marxism Wuhan University of Engineering Science Wuhan 430200 PR China Business School Quanzhou Normal University Quanzhou 362000 PR China

The rise of artificial intelligence (AI) has significantly impacted the creative domain, once considered exclusive to humans, sparking intense academic debate regarding the authenticity of AI creativity and the protection of AI-generated works. This paper critically examines dominant technological culture ideologies across various posthuman societal scenarios and employs the “4P Perspectives of Machine Creativity” framework to assess the legitimacy of AI creativity. It analyzes the dilemmas and anthropocentric constraints embedded in current legal approaches and proposes context-specific protection strategies for AI-generated works. The study finds that the underlying cultural ideologies shaping Industry 4.0+5.0, Cyborg Society, and Technological Singularity scenarios have influenced the evolution of AI ethics and legal interpretations. While the Person, Process, and Product dimensions confirm the existence of AI creativity, legal and cultural norms under the Press dimension remain the main barriers to its recognition. The reluctance to legally protect AI-generated works reflects persistent anthropocentric biases, which may ultimately undermine the interests of human creators. This paper recommends allocating rights within the existing copyright framework based on AI’s “distributed agency” in the Industry 4.0+5.0 scenario, and exploring limited or full legal personhood for AI in more advanced posthuman contexts to ensure appropriate rights attribution.

关键词： AI Posthumanism Anthropocentrism Creativity Legal protection

Factors that affect the technological transition of firms toward the industry 4.0 technologies

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Kim, Seung Hwan Jeon, Jeong hwan Aridi, Anwar Jun, Bogang Technology Management Economics and Policy Program Seoul National University Seoul Korea Republic of Department of Electrical Engineering Ulsan National Institute of Science and Technology Ulsan Korea Republic of Graduate School of Artificial Intelligence Ulsan National Institute of Science and Technology Ulsan Korea Republic of Finance Competitiveness and Innovation Global Practice Europe and Central Asia Unit The World Bank Group Department of Economics Inha University Korea Republic of Department of Data Science Inha University Korea Republic of Research Center for Small Businesses Ecosystem Inha University Korea Republic of

This research aims to identify factors that affect the technological transition of firms toward industry 4.0 technologies (I4Ts) focusing on firm capabilities and policy impact using relatedness and complexity measures. For the analysis, a unique dataset of Korean manufacturing firms’ patent and their financial and market information was used. Following the Principle of Relatedness, which is a recently shaped empirical principle in the field of economic complexity, economic geography, and regional studies, we build a technology space and then trace each firm’s footprint on the space. Using the technology space of firms, we can identify firms that successfully develop a new industry 4.0 technology and examine whether their accumulated capabilities in their previous technology domains positively affect their technological diversification and which factors play a critical role in their transition towards industry 4.0. In addition, by combining data on whether the firms received government support on R&D activities, we further analyzed the role of government policy in supporting firms’ knowledge activity in new industry 4.0 technologies. We found that firms with higher related technologies and more government support are more likely to enter new I4Ts. We expect our research to inform policymakers who aim to diversify firms’ technological capabilities into I4Ts. © 2022, CC BY.

关键词： Industry 4.0

Fine-Grained Garment Parsing: A Body Generation Approach

学校读者我要写书评

暂无评论

Fine-Grained Garment Parsing: A Body Generation Approach

IEEE International Conference on Multimedia and Expo (ICME)

作者： Peng Zhang Yuwei Zhang Shan Huang Zhi Wang Department of Computer Science and Technology Tsinghua University Tsinghua Shenzhen International Graduate School Peng Cheng Laboratory

ISBN: (数字)9781728113319

ISBN: (纸本)9781728113326

Current human parsing methods segment an image into different semantic parts including background, body parts and garments. A major limitation of today's human parsing methodologies is that they are not able to provide fine-grained garment segmentation (e.g., left and right sleeves), and it is mainly due to the lack of a dataset with such fine-grained semantic garment part labels. To tackle this, we propose a body generation approach for fine-grained garment parsing. In particular, we first use a body generation module based on image inpainting, to locate the fine-grained garment parts corresponding to where the generated body parts are, e.g., the left sleeve is assumed to be associated with the left arm; we then extract the garment parts from the original whole garment based on the positions above. In our experiments based on a public dataset focusing on top clothing images, our solution can effectively separate a top garment into a left sleeve, a right sleeve and front, as compared to state-of-the-art solutions that parse it as a whole.

关键词： Clothing Image segmentation Convolution Semantics Task analysis Computational modeling Body regions

Study on an appropriate value of the maximum transmission rate over congested IEEE802.11g Wireless LAN based on WebQoE 8

学校读者我要写书评

暂无评论

Study on an appropriate value of the maximum transmission ra...

8th IEEE Global Conference on Consumer Electronics, GCCE 2019

作者： Hayakawa, Masato Ito, Yoshihiro Graduate School of Engineering Nagoya Institute of Technology Department of Computer Science and Engineering Nagoya466-8555 Japan

ISBN: (纸本)9781728135755

This paper studies an appropriate value of the maximum transmission rate under congested for IEEE802.11g wireless LAN by experiments with subjects. In our experiment, we studied the appropriate value of maximum transmission rate according to Quality of Experience(QoE) for Web services. This paper treats actual two Web services: Google map and Amazon. This paper studied the QoE-based appropriate values of the maximum transmission rate by experiment. As a result, for Google map, the value of the maximum transmission which can maintain high QoE regardless of the degree of congestion exists. However the most suitable maximum transmission rate was changed according to the number of user. On the other, for Amazon, the appropriate value of the maximum transmission rate does not change regardless of the number of user. © 2019 IEEE.

关键词： Web services

CSformer: Bridging Convolution and Transformer for Compressive Sensing

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Ye, Dongjie Ni, Zhangkai Wang, Hanli Zhang, Jian Wang, Shiqi Kwong, Sam The Department of Computer Science City University of Hong Kong 999077 Hong Kong The Department of Computer Science & Technology Key Laboratory of Embedded System and Service Computing Ministry of Education Shanghai Institute of Intelligent Science and Technology Tongji University Shanghai200092 China The School of Electronic and Computer Engineering Peking University Shenzhen Graduate School Shenzhen518055 China The Peng Cheng Laboratory Shenzhen518052 China The City University of Hong Kong Shenzhen Research Institute Shenzhen518057 China

Convolution neural networks (CNNs) have succeeded in compressive image sensing. However, due to the inductive bias of locality and weight sharing, the convolution operations demonstrate the intrinsic limitations in modeling the long-range dependency. Transformer, designed initially as a sequence-to-sequence model, excels at capturing global contexts due to the self-attention-based architectures even though it may be equipped with limited localization abilities. This paper proposes CSformer, a hybrid framework that integrates the advantages of leveraging both detailed spatial information from CNN and the global context provided by transformer for enhanced representation learning. The proposed approach is an end-to-end compressive image sensing method, composed of adaptive sampling and recovery. In the sampling module, images are measured block-by-block by the learned sampling matrix. In the reconstruction stage, the measurement is projected into dual stems. One is the CNN stem for modeling the neighborhood relationships by convolution, and the other is the transformer stem for adopting global self-attention mechanism. The dual branches structure is concurrent, and the local features and global representations are fused under different resolutions to maximize the complementary of features. Furthermore, we explore a progressive strategy and window-based transformer block to reduce the parameter and computational complexity. The experimental results demonstrate the effectiveness of the dedicated transformer-based architecture for compressive sensing, which achieves superior performance compared to state-of-the-art methods on different datasets. Copyright © 2021, The Authors. All rights reserved.

关键词： Compressed sensing

A secure data sharing scheme based on searchable public key encryption for authorized multi-receiver

学校读者我要写书评

暂无评论

Journal of Systems Architecture 2025年 167卷

作者： Huan Li Lunzhi Deng Yating Gu Na Wang Yanli Chen Siwei Li School of Mathematical Sciences Guizhou Normal University Guiyang 550025 China Guizhou Provincial Specialized Key Laboratory of Information Security Technology in Higher Education Institutions Guiyang 550025 China School of Cyber Science and Technology Beihang University Beijing 100191 China School of Big Data and Computer Science Guizhou Normal University Guiyang 550025 China Graduate School Guizhou Normal University Guiyang 550025 China

To ensure data confidentiality, data sharers usually choose to encrypt the data before uploading it to cloud storage. Data sharing is an important way to realize the value of data. Therefore, how to share encrypted data stored in the cloud among authorized users is a pressing issue that needs to be addressed. Public key encryption schemes with keyword search provide an effective solution to this problem. In this paper, we first analyze Yang et al.’ scheme Yang et al. (2023), and point out that the scheme does not realize the indistinguishability of ciphertext and trapdoor. Then, we propose a new data sharing scheme with searchable public key encryption for authorized multi-receiver (SDS-SPKE), which not only realizes the search function, but also realizes the key update, user revocation. Additionally, we provide the security proofs of the scheme, which reveals that our scheme realizes the indistinguishability of ciphertext and trapdoor, and solves the problem of single-key-exposure. Finally, we compare the performance of SDS-SPKE with five other searchable encryption schemes, and the experimental results show that our scheme offers superior efficiency.

关键词：

Highly stable full-color display device with VLC application potential using semipolarμLEDs and all-inorganic encapsulated perovskite nanocrystal

学校读者我要写书评

暂无评论

Photonics Research 2021年第11期9卷 2132-2143页

作者： TINGZHU WU YUE LIN YU-MING HUANG MENG LIU KONTHOUJAM JAMES SINGH WANSHENG LIN TINGWEI LU XI ZHENG JIANYANG ZHOU HAO-CHUNG KUO ZHONG CHEN School of Electronic Science and Engineering Fujian Engineering Research Center for Solid-State LightingXiamen UniversityXiamen 361005China Fujian Science&Technology Innovation Laboratory for Energy Materials of China Xiamen 361005China Department of Photonics and Graduate Institute of Electro-Optical Engineering College of Electrical and Computer EngineeringTaiwan Chiao Tung UniversityHsinchu 30010China Semiconductor Research Center Hon Hai Research InstituteTaipei 11492China

A promising approach for the development of effective full-color displays is to combine blue micro LEDs(μLEDs)with color conversion *** nanocrystals(PNCs)are notable for their tolerance to defects and provide excellent photoluminescence quantum yields and high color purity compared to metal chalcogenide quantum *** stability of PNCs in ambient conditions and under exposure to blue light can be improved using a SiO;*** study proposes a device that could be used for both display and visible light communication(VLC)*** semipolar blueμLED array fabricated in this study shows a negligible wavelength shift,indicating a significant reduction in the quantum confined Stark *** to its shorter carrier lifetime,the semipolar μLED array exhibits an impressive peak 3 dB bandwidth of 655 MHz and a data transmission rate of1.2 Gb/s corresponding to an injection current of 200 m *** PNC-μLED device assembled from a semipolarμLED array with PNCs demonstrates high color stability and wide color-gamut features,achieving 127.23% and 95.00%of the National Television Standards Committee standard and Rec.2020 on the CIE 1931 color diagram,*** results suggest that the proposed PNC-μLED device is suitable for both display-related and VLC applications.

关键词： VLC quantum polar

Enriching and Characterizing T-Cell Repertoires from 3' Barcoded Single-Cell Whole Transcriptome Amplification Products

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Jivanjee, Tasneem Ibrahim, Samira Nyquist, Sarah K. James Gatter, G. Bromley, Joshua D. Jaiswal, Swati Berger, Bonnie Behar, Samuel M. Christopher Love, J. Shalek, Alex K. Institute for Medical Engineering & Science Department of Chemistry Massachusetts Institute of Technology CambridgeMA United States Koch Institute for Integrative Cancer Research Massachusetts Institute of Technology CambridgeMA United States Ragon Institute of MGH MIT and Harvard CambridgeMA United States Broad Institute of MIT and Harvard CambridgeMA United States Massachusetts Institute of Technology CambridgeMA United States Program in Computational and Systems Biology Massachusetts Institute of Technology CambridgeMA United States Computer Science and Artificial Intelligence Laboratory Massachusetts Institute of Technology CambridgeMA United States Microbiology Graduate Program Massachusetts Institute of Technology CambridgeMA02139 United States Department of Microbiology and Physiological Systems University of Massachusetts Medical School WorcesterMA United States Department of Mathematics Massachusetts Institute of Technology CambridgeMA United States Department of Chemical Engineering Massachusetts Institute of Technology CambridgeMA United States

Antigen-specific T cells play an essential role in immunoregulation and many diseases such as cancer. Characterizing the T cell receptor (TCR) sequences that encode T cell specificity is critical for elucidating the antigenic determinants of immunological diseases and designing therapeutic remedies. However, methods of obtaining single-cell TCR sequencing data are labor and cost intensive, typically requiring both cell sorting and full length single-cell RNA-sequencing (scRNA-seq). New high-throughput 3' cell-barcoding scRNA-seq methods can simplify and scale this process;however, they do not routinely capture TCR sequences during library preparation and sequencing. While 5' cell-barcoding scRNA-seq methods can be used to examine TCR repertoire at single-cell resolution, doing so requires specialized reagents which cannot be applied to samples previously processed using 3' cell-barcoding methods. Here, we outline a method for sequencing TCRα and TCRβ transcripts from samples already processed using 3' cell-barcoding scRNA-seq platforms, ensuring TCR recovery at a single-cell resolution. In short, a fraction of the 3' barcoded whole transcriptome amplification (WTA) product typically used to generate a massively parallel 3' scRNA-seq library is enriched for TCR transcripts using biotinylated probes, and further amplified using the same universal primer sequence from WTA. Primer extension using TCR V-region primers and targeted PCR amplification using a second universal primer results in a 3' barcoded single-cell CDR3-enriched library that can be sequenced with custom sequencing primers. Coupled with 3' scRNA-seq of the same WTA, this method enables simultaneous analysis of single-cell transcriptomes and TCR sequences which can help interpret inherent heterogeneity among antigen-specific T cells and salient disease biology. The method presented here can also be adapted readily to enrich and sequence other transcripts of interest from both 3' and 5' barcoded scRNA-seq

关键词： RNA

Generalization error bound for hyperbolic ordinal embedding

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Suzuki, Atsushi Nitanda, Atsushi Wang, Jing Xu, Linchuan Cavazza, Marc Yamanishi, Kenji School of Computing and Mathematical Sciences Faculty of Liberal Arts and Sciences University of Greenwich United Kingdom Department of Artificial Intelligence Faculty of Computer Science and Systems Engineering Kyushu Institute of Technology Japan Department of Computing Hong Kong Polytechnic University Hong Kong Department of Mathematical Informatics Graduate School of Information Science and Technology University of Tokyo Japan

Hyperbolic ordinal embedding (HOE) represents entities as points in hyperbolic space so that they agree as well as possible with given constraints in the formof entity i ismore similar to entity j than to entity k. It has been experimentally shown that HOE can obtain representations of hierarchical data such as a knowledge base and a citation network effectively, owing to hyperbolic space's exponential growth property. However, its theoretical analysis has been limited to ideal noiseless settings, and its generalization error in compensation for hyperbolic space's exponential representation ability has not been guaranteed. The difficulty is that existing generalization error bound derivations for ordinal embedding based on the Gramian matrix do not work in HOE, since hyperbolic space is not inner-product space. In this paper, through our novel characterization of HOE with decomposed Lorentz Gramian matrices, we provide a generalization error bound of HOE for the first time, which is at most exponential with respect to the embedding space's radius. Our comparison between the bounds of HOE and Euclidean ordinal embedding shows that HOE's generalization error is reasonable as a cost for its exponential representation ability. Copyright © 2021, The Authors. All rights reserved.

关键词： Knowledge based systems