检索结果-内蒙古大学图书馆

arXiv 2023年

作者： Wang, Lean Li, Lei Dai, Damai Chen, Deli Zhou, Hao Meng, Fandong Zhou, Jie Sun, Xu National Key Laboratory for Multimedia Information Processing School of Computer Science Peking University China Pattern Recognition Center WeChat AI Tencent Inc. China

In-context learning (ICL) emerges as a promising capability of large language models (LLMs) by providing them with demonstration examples to perform diverse tasks. However, the underlying mechanism of how LLMs learn from the provided context remains under-explored. In this paper, we investigate the working mechanism of ICL through an information flow lens. Our findings reveal that label words in the demonstration examples function as anchors: (1) semantic information aggregates into label word representations during the shallow computation layers' processing;(2) the consolidated information in label words serves as a reference for LLMs' final predictions. Based on these insights, we introduce an anchor re-weighting method to improve ICL performance, a demonstration compression technique to expedite inference, and an analysis framework for diagnosing ICL errors in GPT2-XL. The promising applications of our findings again validate the uncovered ICL working mechanism and pave the way for future studies. © 2023, CC BY.

关键词： Demonstrations

来源：评论

学校读者我要写书评

暂无评论

Detection level of mango based on neural network and digital image

引用

Journal of Chemical and Pharmaceutical Research 2014年第2期6卷 129-132页

作者： Li, Xiaoling Yuan, Jimin School of Computer Science and Technology of Chengdu University China University Key Laboratory of Pattern Recognition and Intelligent Information Processing China Department Computer Engineering of Chengdu Aeronautic Vocational and Technical College China

In view of the draw backs of mango grade identification in China, which still relies on photoelectric sorting and manual separation, this paper presents a processing method on the basis of the technology of computer vision and digital image. Utilizing image processing technology, the researcher calculated the length of the long-short-axis, marked the location of it and calculated the 7 parameters, chroma, length, width and etc.,4 of which are chosen as the key characteristics of the BP input of network to build a network and identify the level of mango through analysis of the external characteristics of mango. The method is based on traditional characteristics detection, using boundary tracking algorithm and the length of the new long-short-axis detection algorithm. The result of experiment indicates that the calculating method and judging of the level of mango are precise and accurate, with an average recognition rate of 92%. Therefore, the method has a great practical value, which can be applied to other agricultural products classification.

关键词： Fruits

来源：评论

学校读者我要写书评

暂无评论

Consensus Control for Second-Order Multi-Agent with Time-Delay on Antagonistic Networks

Consensus Control for Second-Order Multi-Agent with Time-Del...

引用

第35届中国控制会议

作者： Penglin LI Yue LIU Yadong ZHAO Jie YANG Institute of Image Processing and Pattern Recognition Shanghai Jiao Tong University State Key Laboratory of Industrial Control Technology Institute of Industrial Process ControlCollege of Control Science and EngineeringZhejiang University Key Laboratory of System Control and Information Processing Ministry of Education

ISBN: (纸本)9781509009107

In this paper,we discuss consensus problems for antagonistic networks with double integrator *** cases are analyzed:(1) undirected graphs with fixed topology on antagonistic networks;(2) undirected graphs with fixed topology and time-delay on antagonistic *** both cases,distributed consensus protocols are proposed,with sufficient and necessary conditions *** is proved that the largest tolerable time-delay is only related to the largest eigenvalue of the graph ***,simulations are provided to demonstrate the obtained theoretical results.

关键词： Consensus Fixed Topology Multi-agent Antagonistic Networks Time-delay

来源：评论

学校读者我要写书评

暂无评论

Selection of optimal wavelet basis for singularity detection of non-stationary signal

Selection of optimal wavelet basis for singularity detection...

引用

2nd Annual Conference on Electrical and Control Engineering, ICECE 2011

作者： Xu, Licheng Xue, Misheng Huazhong University of Science and Technology Institute for Pattern Recognition and Artificial Intelligence State Key Laboratory for Multi-spectral Information Processing Technology Wuhan 430074 Hubei China

ISBN: (纸本)9781424481637

By using wavelet transform modulus maximum principle for non-stationary signal singularity detection is a kind of very good method. Through to the various wavelet singularity extracted, the analysis results can be divided into four types: accurate location, the approximate location, overlapping effect, rim effect. According to the classification we learn the optimal wavelet basis should has the following features: the optimum wavelet basis should have strong ability of detecting and precision, and at the same time the influence of overlap and rim should be as small as possible. According to these characteristics, a discriminant function is constructed. The wavelet basis makes the largest discriminant function value is optimal. The experimental results show that the method in this paper according to find out the optimum wavelet basis did more than other wavelet detection effect better. © 2011 IEEE.

关键词： Wavelet transforms

来源：评论

学校读者我要写书评

暂无评论

Learning Analysis of Kernel Ridgeless Regression with Asymmetric Kernel Learning

arXiv

引用

arXiv 2024年

作者： He, Fan He, Mingzhen Shi, Lei Huang, Xiaolin Suykens, Johan A.K. STADIUS Center for Dynamical Systems Signal Processing and Data Analytics KU Leuven Leuven Belgium MOE Key Laboratory of System Control and Information Processing Institute of Image Processing and Pattern Recognition Shanghai Jiao Tong University Shanghai China Shanghai Key Laboratory for Contemporary Applied Mathematics School of Mathematical Sciences Fudan University Shanghai200433 China Shanghai Artificial Intelligence Laboratory Shanghai200232 China MOE Key Laboratory of System Control and Information Processing Institute of Image Processing and Pattern Recognition Institute of Medical Robotics Shanghai Jiao Tong University Shanghai200240 China

Ridgeless regression has garnered attention among researchers, particularly in light of the "Benign Overfitting" phenomenon, where models interpolating noisy samples demonstrate robust generalization. However, kernel ridgeless regression does not always perform well due to the lack of flexibility. This paper enhances kernel ridgeless regression with Locally-Adaptive-Bandwidths (LAB) RBF kernels, incorporating kernel learning techniques to improve performance in both experiments and theory. For the first time, we demonstrate that functions learned from LAB RBF kernels belong to an integral space of Reproducible Kernel Hilbert Spaces (RKHSs). Despite the absence of explicit regularization in the proposed model, its optimization is equivalent to solving an 0-regularized problem in the integral space of RKHSs, elucidating the origin of its generalization ability. Taking an approximation analysis viewpoint, we introduce an lq-norm analysis technique (with 0 © 2024, CC BY.

关键词： Vector spaces

来源：评论

学校读者我要写书评

暂无评论

Research on Large-Scale Scene Infrared Simulation Based on Image Classification

Research on Large-Scale Scene Infrared Simulation Based on I...

引用

International Workshop on Multi-Platform/Multi-Sensor Remote Sensing and Mapping

作者： Lamei Zou Tianxu Zhang Wei Lv Guoqing Ao Institute of Pattern Recognition and Artificial Intelligence National key laboratory of science and technology on multi-spectral information processing technologies Huazhong University of Science and Technology Wuhan China

Vega has been widely used in Virtual Reality (VR) field. Vega infrared (IR) module can implement IR simulation, but Vega IR imaging simulation's general approach does not apply to large-scale scene's infrared simulation problem. This article deeps into large-scale scene IR image simulation through Vega infrared module based on visible image, the scene's corresponding Digital Elevation Model (DEM) data and targets' 3D model. A real time large-scale IR simulation system is successfully designed and realized in this paper. At the aspect of image classification, a coarse to fine k-means clustering method based on the consistency of image color is proposed. The proposed automatic texture material mapping method based on the development of Vega TMM tool provides a fast and convenient way for large-scale scene's IR simulation based on Vega IR module. And finally, results of the large-scale scene infrared simulation show the effectiveness of this method.

关键词： Solid modeling Image color analysis Materials Atmospheric modeling Image classification Three dimensional displays Pixel

来源：评论

学校读者我要写书评

暂无评论

A Decentralized Framework for Kernel PCA with Projection Consensus Constraints

arXiv

引用

arXiv 2022年

作者： He, Fan Yang, Ruikai Shi, Lei Huang, Xiaolin Institute of Image Processing and Pattern Recognition Shanghai Jiao Tong University The MOE Key Laboratory of System Control and Information Processing Shanghai200240 China Shanghai Key Laboratory for Contemporary Applied Mathematics China School of Mathematical Sciences Fudan University Shanghai200240 China

This paper studies kernel PCA in a decentralized setting, where data are distributively observed with full features in local nodes and a fusion center is prohibited. Compared with linear PCA, the use of kernel brings challenges to the design of decentralized consensus optimization: the local projection directions are data-dependent. As a result, the consensus constraint in distributed linear PCA is no longer valid. To overcome this problem, we propose a projection consensus constraint and obtain an effective decentralized consensus framework, where local solutions are expected to be the projection of the global solution on the column space of local dataset. We also derive a fully non-parametric, fast and convergent algorithm based on alternative direction method of multiplier, of which each iteration is analytic and communication-efficient. Experiments on a truly parallel architecture are conducted on real-world data, showing that the proposed decentralized algorithm is effective to utilize information of other nodes and takes great advantages in running time over the central kernel PCA. Copyright © 2022, The Authors. All rights reserved.

关键词： Principal component analysis

来源：评论

学校读者我要写书评

暂无评论

Fast Marine Route Planning for UAV Using Improved Sparse A* Algorithm

Fast Marine Route Planning for UAV Using Improved Sparse A* ...

引用

International Conference on Genetic and Evolutionary Computing (ICGEC)

作者： Xin Yang Ming-yue Ding Cheng-ping Zhou Institute for Pattern Recognition and Artificial Intelligence (IPRAI) Image Processing and Intelligence Control Key Laboratory of Education Ministry of China Huazhong University of Science and Technology Wuhan Hubei China School of Life Science and Technology Image Processing and Intelligence Control Key Laboratory of Education Ministry of China Huazhong University of Science and Technology Wuhan Hubei China Institute for Pattern Recognition and Artificial Intelligence (IPRAI) State Key Laboratory for Multispectral Information Processing Technologies Huazhong University of Science and Technology Wuhan Hubei China

This paper focuses on route planning, especially for unmanned aircrafts in marine environment. Firstly, new heuristic information is adopted such as threat-zone, turn maneuver and forbid-zone based on voyage heuristic information. Then, the cost function is normalized to obtain more flexible and reasonable routes. Finally, an improved sparse A* search algorithm is employed to enhance the planning efficiency and reduce the planning time. Experiment results showed that the improved algorithm for aircraft in maritime environment could find a combinational optimum route quickly, which detoured threat-zones, with fewer turn maneuver, totally avoiding forbid-zones, and shorter voyage.

关键词： Unmanned aerial vehicles Planning Synthetic aperture sonar Cost function Algorithm design and analysis Heuristic algorithms Leg

来源：评论

学校读者我要写书评

暂无评论

One-Class SVM assisted accurate tracking

One-Class SVM assisted accurate tracking

引用

ACM/IEEE International Conference on Distributed Smart Cameras, ICDSC

作者： Keren Fu Chen Gong Yu Qiao Jie Yang Irene Guy Institute of Image Processing and Pattern Recognition Shanghai Jiao Tong University China Ministry of Education of China Key Laboratory of System Control and Information Processing Shanghai China Chalmers tekniska hogskola Goteborg SE

Recently, tracking is regarded as a binary classification problem by discriminative tracking methods. However, such binary classification may not fully handle the outliers, which may cause drifting. In this paper, we argue that tracking may be regarded as one-class problem, which avoids gathering limited negative samples for background description. Inspired by the fact the positive feature space generated by One-Class SVM is bounded by a closed sphere, we propose a novel tracking method utilizing One-Class SVMs that adopt HOG and 2bit-BP as features, called One-Class SVM Tracker (OCST). Simultaneously an efficient initialization and online updating scheme is also proposed. Extensive experimental results prove that OCST outperforms some state-of-the-art discriminative tracking methods on providing accurate tracking and alleviating serious drifting.

关键词： Target tracking Support vector machines Training Feature extraction Face Boosting Semisupervised learning

来源：评论

学校读者我要写书评

暂无评论

Towards multimodal response generation with exemplar augmentation and curriculum optimization

arXiv

引用

arXiv 2020年

作者： Lei, Zeyang Li, Zekang Zhang, Jinchao Meng, Fandong Feng, Yang Yang, Yujiu Niu, Cheng Zhou, Jie Tsinghua University China Key Laboratory of Intelligent Information Processing Institute of Computing Technology Chinese Academy of Sciences Pattern Recognition Center WeChat AI Tencent Inc China

Recently, variational auto-encoder (VAE) based approaches have made impressive progress on improving the diversity of generated responses. However, these methods usually suffer the cost of decreased relevance accompanied by diversity improvements (Zhang et al., 2018b). In this paper, we propose a novel multimodal1 response generation framework with exemplar augmentation and curriculum optimization to enhance relevance and diversity of generated responses. First, unlike existing VAE-based models that usually approximate a simple Gaussian posterior distribution, we present a Gaussian mixture posterior distribution (i.e, multimodal) to further boost response diversity, which helps capture complex semantics of responses. Then, to ensure that relevance does not decrease while diversity increases, we fully exploit similar examples (exemplars) retrieved from the training data into posterior distribution modeling to augment response relevance. Furthermore, to facilitate the convergence of Gaussian mixture prior and posterior distributions, we devise a curriculum optimization strategy to progressively train the model under multiple training criteria from easy to hard. Experimental results on widely used SwitchBoard and DailyDialog datasets demonstrate that our model achieves significant improvements compared to strong baselines in terms of diversity and relevance. Copyright © 2020, The Authors. All rights reserved.

关键词： Curricula

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：