Neural Bellman-Ford Networks: A General Graph Neural Network Framework for Link Prediction 【NeurIPS 2021】神经Bellman-Ford网络:用于链路预测的一般图神经网络框架 链接预测是图的一项非常基础的任务。在传统路径学习方法的启发下，本文提出了一种通用的、灵活的基于路径的链接预测表示学习框架。具体来说，我们将节点对的表示定义为所有路径表示的广义和，每个路径表示都是路径中各边表示的广义乘积。受求解最短路径问题的Bellman-Ford算法的启发，我们证明了所提出的路径公式可以被广义Bellman-Ford算法有效地求解。为了进一步提高路径表示的能力，我们提出了神经BellmanFord网络(NBFNet)，这是一个通用的图神经网络框架，用于解决广义Bellman-Ford算法中使用学习算子的路径表示。NBFNet将广义Bellman-Ford算法参数化，采用3个神经单元，分别对应边界条件、乘法算子和求和算子。NBFNet是非常通用的，涵盖了许多传统的基于路径的方法，并且可以应用于同构图和多关系图(例如，知识图)在转换和归纳设置。在同构图和知识图谱上的实验表明，所提出的NBFNet在转导和归纳设置方面都大大优于现有方法，取得了最新的研究结果。

This book provides a thorough introduction to the formal foundations and practical applications of Bayesian networks. It provides an extensive discussion of techniques for building Bayesian networks that model real-world situations, including techniques for synthesizing models from design, learning models from data, and debugging models using sensitivity analysis. It also treats exact and approximate inference algorithms at both theoretical and practical levels. The author assumes very little background on the covered subjects, supplying in-depth discussions for theoretically inclined readers and enough practical details to provide an algorithmic cookbook for the system developer.

neural-networks-and-deep-learning.pdf

斯宾 在 python 中实现 Sum-Product Networks (SPN) 并提供一些例程来进行推理和学习。 概述 实施和 SPN-BTB，如下所示： A. Vergari、N. Di Mauro 和 F. 埃斯波西托在 ECML-PKDD 2015 上简化、正则化和加强和积网络结构学习。 要求 spyn建立在 ， ， ， ， 和。 用法 data/文件夹中提供了几个数据集。 要运行算法及其网格搜索，请检查bin/文件夹中的脚本。 要从nltcs数据的训练集部分学习单个 SPN，您可以调用： ipython -- bin/learnspn_exp.py nltcs 要获得可能参数的概述，请使用-h ： -h, --help show this help message and exit -k [N_ROW_CLUSTERS], --n-row

Power Distribution Networks with On-Chip Decoupling Capacitors, 2nd edition is dedicated to distributing power in high speed, high complexity integrated circuits with power levels exceeding many tens of watts and power supplies below a volt. This book provides a broad and cohesive treatment of power distribution systems and related design problems, including both circuit network models and design techniques for on-chip decoupling capacitors. The book provides insight and intuition into the behavior and design of on-chip power distribution systems. This book has four primary objectives. The first objective is to describe the impedance characteristics of the overall power distribution system, from the voltage regulator through the printed circuit board and package onto the integrated circuit to the terminals of the on-chip circuitry. The second objective is to discuss the inductive characteristics of on-chip power distribution grids and the related circuit behavior of these structures. The third objective is to present design methodologies for efficiently placing on-chip decoupling capacitors in nanoscale integrated circuits. Finally, the fourth objective is to suggest novel architectures for distributing power across an integrated circuit, as well as provide new methodologies to efficiently analyze and design on-chip power grids. Organized into subareas to provide a more intuitive flow to the reader, this edition adds more than a hundred pages of new content, including inductance models for interdigitated structures, design strategies for multi-layer power grids, advanced methods for efficient power grid design and analysis, and methodologies for simultaneously placing on-chip multiple power supplies and decoupling capacitors. The emphasis of this additional material is on managing the complexity of on-chip power distribution networks.

基于NNAEC-神经网络的声学回声消除：基于NNAEC-神经网络的声学回声消除

Tensorflow中的指针网络 TensorFlow实现。 支持多线程数据管道以减少I / O延迟。 要求 Python 2.7 用法 训练模型： $ python main.py --task=tsp --max_data_length=20 --hidden_dim=512 # download dataset used in the paper $ python main.py --task=tsp --max_data_length=10 --hidden_dim=128 # generate dataset itself 训练模型： $ python main.py $

A very good textbook for researchers working on the NN theory, yet if you just want to know what is NN, it perhaps is too much for you to digest!

PyTorch-NEAT NEAT（增强拓扑的神经进化）方法的PyTorch实现，最初是由Kenneth O. Stanley创建的，是进化神经网络的一种有原则的方法。 。 实验 PyTorch-NEAT当前包含三个内置实验：XOR，单极平衡和汽车爬山。 异或实验 使用以下命令运行： python xor_run.py将运行多达150代，初始种群为150个基因组。 当/如果找到解决方案，将显示解决方案网络以及有关试验的统计信息。 随意运行多个试用版-只需​​增加xor_run.py文件中外部for循环的范围即可。 单极平衡 使用以下命令运行： python pole_run.py将运行多达150代，初始种群为150个基因组。 在OpenAI体育馆环境中跑步。 当/如果找到解决方案，则将在OpenAI体育馆中显示解决方案网络以及评估结果。 汽车登山实验 使用以下命令运行： python m

演示代码（请参阅jupyter笔记本）： 使用深度卷积自动编码器对地震信号进行非监督（自我监督）区分 您可以从这里获取论文： 连结1： 连结2： 您可以从此处获取训练数据集： 参考： Mousavi, S. M., W. Zhu, W. Ellsworth, G. Beroza (2019). Unsupervised Clustering of Seismic Signals Using Deep Convolutional Autoencoders, IEEE Geoscience and Remote Sensing Letters, 1 - 5, doi:10.1109/LGRS.2019.2909218.