RNN时间序列异常检测 在Pytorch中实现的基于RNN的时间序列异常检测器模型。 这是基于RNN的时间序列异常检测器的一种实现，它由时间序列预测和异常分数计算的两阶段策略组成。 要求 Ubuntu 16.04+（在Windows 10上报告了错误。请参阅。欢迎提出建议。） Python 3.5+ 火炬0.4.0+ 脾气暴躁的 Matplotlib Scikit学习 数据集 1.纽约市出租车乘客人数 提供的纽约市出租车乘客数据流 Cui，Yuwei等人进行了预处理（以30分钟为间隔汇总）。 在中 ， 2.心电图（ECG） ECG数据集包含对应于心室前收缩的单个异常 3. 2D手势（视频监控） 视频中手势的XY坐标 4.呼吸 一个病人的呼吸（通过胸廓扩展测量，采样率10Hz） 5.航天飞机 航天飞机Marotta阀的时间序列 6.电力需求 荷兰研究机构一年的电力需求 时

pro_gan_pytorch 包包含 ProGAN 的实现。 论文题为“渐进式增长的 GAN 以提高质量、稳定性和变化”。 链接 -> 训练示例 -> :star: [新] 预训练模型： 请找下预训练模型saved_models/在目录 :star: [新]演示： 存储库现在在samples/目录下包含一个潜在空间插值动画演示。 只需从上面提到的 drive_link 下载所有预训练的权重，并将它们放在demo.py脚本旁边的samples/目录中。 请注意，在demo.py脚本的开头有一些demo.py参数，以便您可以使用它。 该演示加载随机点的图像，然后在它们之间进行线性插值以生成平滑的动画。 你需要有一个好的 GPU（至少 GTX 1070）才能在演示中看到强大的 FPS。 然而，可以优化演示以并行生成图像（目前它是完全顺序的）。 为了在 Generator 中加载权重，该过程是 P

Neural Network with Extended Kalman Filter

ACL Anthology Network衍生产品 aanrelease2013.tar.gz一团糟。 编码无处不在。 某些数据显然丢失或存在，但已损坏。 尚不清楚哪些文件是原始数据，哪些是从这些数据派生的统计信息。 看起来像它们应该是脚本输出的文件实际上是这些脚本的帮助消息。 许多papers_text/文件（从PDF提取的纯文本）存在以下问题： 由于不兼容的PDF编码（例如P00-1032 ， W06-3709 ），其中一些显得乱七八糟 由于原始PDF中的OCR错误（从有问题的T75-2033到无法使用的J79-1013 ），其他文件在不同程度上都乱七八糟。 其他人不是英语（例如C73-2029 ） 其他包含完全不同的论文的文本（例如L08-1302 ） Makefile声明性地提供了有关问题和涉及的清理工作的一些文档。 该存储库不包含任何原始数据，仅包含有关如何修

elman_neural_networks：尝试使用Matlab的神经网络工具箱（由https：//crl.ucsd.edu~elmanPapersfsit.pdf撰写）题为“及时发现结构”的Jeff Elman的作品。

optical network

Written by networking veteran with 20 years of experience, Network Warrior provides a thorough and practical introduction to the entire network infrastructure, from cabling to the routers. What you need to learn to pass a Cisco certification exam such as CCNA and what you need to know to survive in the real world are two very different things. The strategies that this book offers weren 't on the exam, but they 're exactly what you need to do your job well.

Computer Networks: A Systems Approach, Fifth Edition, explores the key principles of computer networking, with examples drawn from the real world of network and protocol design. Using the Internet as the primary example, this best-selling and classic textbook explains various protocols and networking technologies. The systems-oriented approach encourages students to think about how individual network components fit into a larger, complex system of interactions. This book has a completely updated content with expanded coverage of the topics of utmost importance to networking professionals and students, including P2P, wireless, network security, and network applications such as e-mail and the Web, IP telephony and video streaming, and peer-to-peer file sharing. There is now increased focus on application layer issues where innovative and exciting research and design is currently the center of attention. Other topics include network design and architecture; the ways users can connect to a network; the concepts of switching, routing, and internetworking; end-to-end protocols; congestion control and resource allocation; and end-to-end data. Each chapter includes a problem statement, which introduces issues to be examined; shaded sidebars that elaborate on a topic or introduce a related advanced topic; What’s Next? discussions that deal with emerging issues in research, the commercial world, or society; and exercises. This book is written for graduate or upper-division undergraduate classes in computer networking. It will also be useful for industry professionals retraining for network-related assignments, as well as for network practitioners seeking to understand the workings of network protocols and the big picture of networking. Completely updated content with expanded coverage of the topics of utmost importance to networking professionals and students, including P2P, wireless, security, and applications Increased focus on application layer issues where innovative and exciting research and design is currently the center of attention Free downloadable network simulation software and lab experiments manual available

吉林大学软件学院 计算机网络课程设计 套接字编程 VC

很不错的网络安全漏洞扫描系统，内包含详细的源码实现，C++实现