检索结果-内蒙古大学图书馆

SELF-SUPERVISED VIDEO REPRESENTATION LEARNING WITH MOTION-CONTRASTIVE PERCEPTION

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Liu, Jinyu Cheng, Ying Zhang, Yuejie Zhao, Rui-Wei Feng, Rui School of Computer Science Shanghai Collaborative Innovation Center of Intelligent Visual Computing Fudan University China Academy for Engineering and Technology Fudan University China

Visual-only self-supervised learning has achieved significant improvement in video representation learning. Existing related methods encourage models to learn video representations by utilizing contrastive learning or designing specific pretext tasks. However, some models are likely to focus on the background, which is unimportant for learning video representations. To alleviate this problem, we propose a new view called long-range residual frame to obtain more motion-specific information. Based on this, we propose the Motion-Contrastive Perception Network (MCPNet), which consists of two branches, namely, Motion Information Perception (MIP) and Contrastive Instance Perception (CIP), to learn generic video representations by focusing on the changing areas in videos. Specifically, the MIP branch aims to learn fine-grained motion features, and the CIP branch performs contrastive learning to learn overall semantics information for each instance. Experiments on two benchmark datasets UCF-101 and HMDB-51 show that our method outperforms current state-of-the-art visual-only self-supervised approaches. Copyright © 2022, The Authors. All rights reserved.

关键词： Video recording

LVOS: A Benchmark for Long-term Video Object Segmentation

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Hong, Lingyi Chen, Wenchao Liu, Zhongying Zhang, Wei Guo, Pinxue Chen, Zhaoyu Zhang, Wenqiang School of Computer Science Fudan University China Shanghai Key Laboratory of Intelligent Information Processing China Academy for Engineering and Technology Fudan University China

Existing video object segmentation (VOS) benchmarks focus on short-term videos which just last about 3-5 seconds and where objects are visible most of the time. These videos are poorly representative of practical applications, and the absence of long-term datasets restricts further investigation of VOS on the application in realistic scenarios. So, in this paper, we present a new benchmark dataset and evaluation methodology named LVOS, which consists of 220 videos with a total duration of 421 minutes. To the best of our knowledge, LVOS is the first densely annotated long-term VOS dataset. The videos in our LVOS last 1.59 minutes on average, which is 20 times longer than videos in existing VOS datasets. Each video includes various attributes, especially challenges deriving from the wild, such as long-term reappearing and cross-temporal similar objeccts. Moreover, we provide additional language descriptions to encourage the exploration of integrating linguistic and visual features for video object segmentation. Based on LVOS, we assess existing video object segmentation algorithms and propose a Diverse Dynamic Memory network (DDMemory) that consists of three complementary memory banks to exploit temporal information adequately. The experiment results demonstrate the strength and weaknesses of prior methods, pointing promising directions for further study. Our objective is to provide the community with a large and varied benchmark to boost the advancement of long-term VOS. Data and code are available at https:// ***/***/. Copyright © 2022, The Authors. All rights reserved.

关键词： Visual languages

High Quality Factor AlScN Lamb Wave Resonators Using NbN/Al Top Electrodes at Cryogenic Temperature

学校读者我要写书评

暂无评论

High Quality Factor AlScN Lamb Wave Resonators Using NbN/Al ...

International Conference on Solid State Sensors and Actuators (TRANSDUCERS)

作者： Zhifang Luo Shuai Shao Peng Dong Haowen Guo Xinbo Zou Jun Li Chengkuo Lee Tao Wu School of Information Science and Technology ShanghaiTech University China Shanghai Institute of Microsystem and Information Technology Chinese Academy of Sciences China University of Chinese Academy of Sciences China Department of Electrical & Computer Engineering National University of Singapore Singapore Center for Intelligent Sensors and MEMS National University of Singapore Singapore School of Physical Science and Technology ShanghaiTech University China Shanghai Engineering Research Center of Energy Efficient and Custom AI IC China

Coupling between phonons and qubits provides a solid quantum platform for information transmission and processing. We report the high quality factor (Q a ) aluminum scandium nitride (AlScN) Lamb wave resonators (LWRs) at cryogenic temperature for quantum applications. Superconducting 60 nm NbN/30nm Al serves as top electrodes for ultra-low energy dissipation at around 11 K. Al 0.84 Sc 0.16 N provides a high piezoelectric constant for high electro-mechanical coupling $\left( {{{\text{k}}_{\text{t}}}^{\text{2}}} \right)$ of resonators. Additionally, LWRs allow lithography-defined resonant frequency for flexible multi-frequency phonon sources. The fabricated LWR shows Q a over 6000, and resulting in a record figure-of-merit $\left( {{\text{FoM }} = {\text{ }}{{\text{k}}_{\text{t}}}^{\text{2}}{\text{*}}{{\text{Q}}_{\text{a}}}} \right)$ of 211 for AlScN resonators.

关键词：

When Neural Code Completion Models Size up the Situation: Attaining Cheaper and Faster Completion through Dynamic Model Inference

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Sun, Zhensu Du, Xiaoning Song, Fu Wang, Shangwen Li, Li Monash University MelbourneVIC Australia State Key Laboratory of Computer Science Institute of Software Chinese Academy of Sciences Beijing China National University of Defense Technology Changsha China Beihang University Yunnan Key Laboratory of Software Engineering Beijing China

Leveraging recent advancements in large language models, modern neural code completion models have demonstrated the capability to generate highly accurate code suggestions. However, their massive size poses challenges in terms of computational costs and environmental impact, hindering their widespread adoption in practical scenarios. Dynamic inference emerges as a promising solution, as it allocates minimal computation during inference while maintaining the model’s performance. In this research, we explore dynamic inference within the context of code completion. Initially, we conducted an empirical investigation on GPT-2, focusing on the inference capabilities of intermediate layers for code completion. We found that 54.4% of tokens can be accurately generated using just the first layer, signifying significant computational savings potential. Moreover, despite using all layers, the model still fails to predict 14.5% of tokens correctly, and the subsequent completions continued from them are rarely considered helpful, with only a 4.2% Acceptance Rate. These findings motivate our exploration of dynamic inference in code completion and inspire us to enhance it with a decision-making mechanism that stops the generation of incorrect code. We thus propose a novel dynamic inference method specifically tailored for code completion models. This method aims not only to produce correct predictions with largely reduced computation but also to prevent incorrect predictions proactively. Our extensive evaluation shows that it can averagely skip 1.7 layers out of 16 layers in the models, leading to an 11.2% speedup with only a marginal 1.1% reduction in ROUGE-L. Copyright © 2024, The Authors. All rights reserved.

关键词： Forecasting

Research on Key Word Information Retrieval Based on Inverted Index 8th

学校读者我要写书评

暂无评论

Research on Key Word Information Retrieval Based on Inverted...

8th International Conference on Artificial Intelligence and Security , ICAIS 2022

作者： Qi, Meihan Fang, Wei Zhao, Yongming Sha, Yu Sheng, Victor S. School of Computer and Software Engineering Research Center of Digital Forensics Ministry of Education Nanjing University of Information Science and Technology Nanjing210044 China State Key Laboratory of Severe Weather Chinese Academy of Meteorological Sciences Beijing100081 China China Meteorological Administration Training Center Beijing China Department of Computer Texas Tech University LubbockTX79409 United States

ISBN: (纸本)9783031067600

With the advent of the era of big data, data has penetrated into every aspect of social life and become an important production factor in various industries. However, while providing convenience to our life, massive information also brings difficulties to information retrieval, which has problems of low retrieval efficiency and poor retrieval accuracy. Therefore, how to improve the efficiency and accuracy of information retrieval has become a key technical problem to be solved. In this paper, we propose a key word information retrieval scheme based on inverted index, which uses the inverted index segmentation algorithm based on keywords to realize the block retrieval of information, improves the efficiency of information retrieval, and improves the accuracy of retrieval content through reasonable error correction of keywords. The results of evaluation indicate that compared with the mainstream information retrieval methods, the keyword information retrieval scheme based on inverted index can carry out faster and more accurate information retrieval. Experimental results show that the keyword information retrieval scheme with inverted index can further improve the efficiency and accuracy of information retrieval. © 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.

关键词： Information retrieval

Enhancing Question Answering for Enterprise Knowledge Bases using Large Language Models

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Jiang, Feihu Qin, Chuan Yao, Kaichun Fang, Chuyu Zhuang, Fuzhen Zhu, Hengshu Xiong, Hui University of Science and Technology of China China Career Science Lab BOSS Zhipin Beijing China PBC School of Finance Tsinghua University Beijing China Institute of Software Chinese Academy of Sciences China Institute of Artificial Intelligence Beihang University China Zhongguancun Laboratory China The Thrust of Artificial Intelligence The Hong Kong University of Science and Technology Guangzhou China The Department of Computer Science and Engineering The Hong Kong University of Science and Technology Hong Kong Guangzhou HKUST Fok Ying Tung Research Institute China

Efficient knowledge management plays a pivotal role in augmenting both the operational efficiency and the innovative capacity of businesses and organizations. By indexing knowledge through vectorization, a variety of knowledge retrieval methods have emerged, significantly enhancing the efficacy of knowledge management systems. Recently, the rapid advancements in generative natural language processing technologies paved the way for generating precise and coherent answers after retrieving relevant documents tailored to user queries. However, for enterprise knowledge bases, assembling extensive training data from scratch for knowledge retrieval and generation is a formidable challenge due to the privacy and security policies of private data, frequently entailing substantial costs. To address the challenge above, in this paper, we propose EKRG, a novel Retrieval-Generation framework based on large language models (LLMs), expertly designed to enable question-answering for Enterprise Knowledge bases with limited annotation costs . Specifically, for the retrieval process, we first introduce an instruction-tuning method using an LLM to generate sufficient document-question pairs for training a knowledge retriever. This method, through carefully designed instructions, efficiently generates diverse questions for enterprise knowledge bases, encompassing both fact-oriented and solution-oriented knowledge. Additionally, we develop a relevance-aware teacher-student learning strategy to further enhance the efficiency of the training process. For the generation process, we propose a novel chain of thought (CoT) based fine-tuning method to empower the LLM-based generator to adeptly respond to user questions using retrieved documents. Finally, extensive experiments on real-world datasets have demonstrated the effectiveness of our proposed framework. © 2024, CC BY.

关键词： Knowledge management

Multi-AMP: Detecting the antimicrobial peptides and their activities using the multi-task learning

学校读者我要写书评

暂无评论

Multi-AMP: Detecting the antimicrobial peptides and their ac...

2021 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2021

作者： Qiaozhen, Meng Tang, Jijun Guo, Fei Tianjin University College of Intelligence and Computing Tianjin China Shenzhen Institute of Advanced Technology Chinese Academy of Sciences Shenzhen China Central South University School of Computer Science and Engineering Changsha China

ISBN: (纸本)9781665401265

Recently due to the broad-spectrum and high-efficiency antibacterial activity, antimicrobial peptides (AMPs) have become the best alternative to antibiotics. With the rapid increase of the antibacterial peptides, many computational methods have been developed to identify the AMPs and their specific antibacterial activities. However, most existing methods regard these two problems as independent sub-problems and ignore the correlation between tasks. In this paper, we propose a method, Multi-AMP, which utilizes multi-task learning and solves two tasks simultaneously: 1) whether a given peptide is AMP, 2) which activities it performs. The two tasks share the parameters at the bottom layers of the model and learn the specific information at the top layers. Experiments indicate that our multi-task model performs better than single-task models and two existing predictors, which can give insights to the drug discovery process. © 2021 IEEE.

关键词： Peptides

Broadband MIMO Antenna with Self-Complementary Resonating Structures for 5G Smartphones 3

学校读者我要写书评

暂无评论

Broadband MIMO Antenna with Self-Complementary Resonating St...

3rd IEEE International Conference on Electronic engineering, ICEEM 2023

作者： Islam, Tanvir Taher, Fatma Parchin, Naser Ojaroudi Amar, Ahmed S. I. Abo Sree, Mohamed Fathy University of Houston Department of Electrical and Computer Engineering HoustonTX77204 United States Zayed University College of Technological Innovation DubaiP.0. Box 19282 United Arab Emirates Edinburgh Napier University School of Computing Engineering and the Built Environment Edinburgh United Kingdom Alexandria University Air Defense College Electronics Department of Electronics and Communication Alexandria Egypt Electronics Research Institute Microstrip Department Cairo Egypt Arab Academy for Science Technology and Maritime Transport Department of Electronics and Communications Engineering Cairo451913 Egypt

ISBN: (纸本)9798350323511

This paper presents an innovative MIMO antenna design using self-complementary resonators with highly desirable characteristics. The MIMO array consists of eight L-shaped slits etched in the ground and fed by L-shaped microstrip lines on the top layers. The operating frequency range for these elements is 2.6-3.6 GHz, falling within the sub-6 GHz 5G cellular communication bands. The MIMO array's compact footprint makes it suitable for integration into smartphones, occupying a relatively small area on the device's board. Despite its size, the design achieves excellent performance in crucial aspects such as S-parameters, radiation patterns, impedance bandwidth, radiation diversity, efficiency, and gain. These results meet the requirements for future 5G handheld smartphones, making the suggested MIMO antenna design highly promising for practical applications in the rapidly advancing 5G technology landscape. © 2023 IEEE.

关键词： MIMO systems

Noncontact Doppler Radar-based Heart Rate Detection on the SVD and ANC

学校读者我要写书评

暂无评论

Noncontact Doppler Radar-based Heart Rate Detection on the S...

2022 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2022

作者： Shi, Qiuxia Hu, Bin Tian, Fuze Zhao, Qinglin Gansu Provincial Key Laboratory of Wearable Computing School of Information Science and Engineering Lanzhou University China School of Computer and Communication Lanzhou University of Technology China Key Laboratory of Special Functional Materials and Structural Design Lanzhou University Ministry of Education Lanzhou China CAS Center for Excellence in Brain Science and Intelligence Technology Shanghai Institutes for Biological Sciences Chinese Academy of Sciences China Joint Research Center for Cognitive Neurosensor Technology of Lanzhou University Institute of Semiconductors Chinese Academy of Sciences China Engineering Research Center of Open Source Software and Real-Time System Lanzhou University Ministry of Education China

ISBN: (纸本)9781665468190

In the Doppler biological radar-based applications of noncontact measurement of vital signs, effectively extracting heartbeat information from weak thoracic mechanical motion is an important problem to be solved. This study is aimed at extracting heartbeat signal via the technology combined with Short Time Fourier Transform (STFT), Singular Value Decomposition (SVD) and Adaptive Noise Canceller (ANC) from radar recording. The simulated data and the data collected by Doppler radar biosensor realized in laboratory are employed to validate the proposed method. The results show that the proposed method has the ability of detection for the heart rate and heart rate variability indexes in rest state, it has certain advantages in time-consuming and detection accuracy. Therefore, the current method provides another way to process vital sign signals recorded by Doppler radar. © 2022 IEEE.

关键词： Singular value decomposition