复旦大学邱锡鹏老师在 Github 上开放的深度学习书籍

https://nndl.github.io/

https://github.com/nndl/nndl.github.io

本书全面介绍了数据挖掘的理论和方法，旨在为读者提供将数据挖掘应用于实际问题所必需的知识。本书涵盖五个主题：数据、分类、关联分析、聚类和异常检测。除异常检测外，每个主题都包含两章：前面一章讲述基本概念、代表性算法和评估技术，后面一章较深入地讨论高级概念和算法。目的是使读者在透彻地理解数据挖掘基础的同时，还能了解更多重要的高级主题。此外，书中还提供了大量示例、图表和习题。 本书适合作为相关专业高年级本科生和研究生数据挖掘课程的教材，同时也可作为数据挖掘研究和应用开发人员的参考书。

In response to the exponentially increasing need to analyze vast amounts of data, Neural Networks for Applied Sciences and Engineering: From Fundamentals to Complex Pattern Recognition provides scientists with a simple but systematic introduction to neural networks. Beginning with an introductory discussion on the role of neural networks in scientific data analysis, this book provides a solid foundation of basic neural network concepts. It contains an overview of neural network architectures for practical data analysis followed by extensive step-by-step coverage on linear networks, as well as, multi-layer perceptron for nonlinear prediction and classification explaining all stages of processing and model development illustrated through practical examples and case studies. Later chapters present an extensive coverage on Self Organizing Maps for nonlinear data clustering, recurrent networks for linear nonlinear time series forecasting, and other network types suitable for scientific data analysis. With an easy to understand format using extensive graphical illustrations and multidisciplinary scientific context, this book fills the gap in the market for neural networks for multi-dimensional scientific data, and relates neural networks to statistics. Features x Explains neural networks in a multi-disciplinary context x Uses extensive graphical illustrations to explain complex mathematical concepts for quick and easy understanding ? Examines in-depth neural networks for linear and nonlinear prediction, classification, clustering and forecasting x Illustrates all stages of model development and interpretation of results, including data preprocessing, data dimensionality reduction, input selection, model development and validation, model uncertainty assessment, sensitivity analyses on inputs, errors and model parameters Sandhya Samarasinghe obtained her MSc in Mechanical Engineering from Lumumba University in Russia and an MS and PhD in Engineering from Virginia Tech, USA. Her neural networks research focuses on theoretical understanding and advancements as well as practical implementations.

Theoretical results suggest that in order to learn the kind of complicated functions that can represent high-level abstractions (e.g., in vision, language, and other AI-level tasks), one may need deep architectures. Deep architectures are composed of multiple levels of non-linear operations, such as in neural nets with many hidden layers or in complicated propositional formulae re-using many sub-formulae. Searching the parameter space of deep architectures is a difficult task, but learning algorithms such as those for Deep Belief Networks have recently been proposed to tackle this problem with notable success, beating the stateof- the-art in certain areas. This monograph discusses the motivations and principles regarding learning algorithms for deep architectures, in particular those exploiting as building blocks unsupervised learning of single-layer models such as Restricted Boltzmann Machines, used to construct deeper models such as Deep Belief Networks.

《神经网络在应用科学与工程中的应用:从基本原理到复杂的模式识别》为读者提供了神经网络方面简单但却系统的介绍。 《神经网络在应用科学和工程中的应用从基本原理到复杂的模式识别》以神经网络在科学数据分析中所扮演角色的介绍性讨论作为开始，给出了神经网络的基本概念。《神经网络在应用科学和工程中的应用从基本原理到复杂的模式识别》首先对用于实际数据分析的神经网络结构进行了综合概述，继而对线性网络进行了大量的介绍，并介绍了所有处理阶段的用于非线性预报和分类的多层感知器。此外，还通过实际例子和个案研究阐述了模型开发技术。后面章节又提出了用于非线性数据聚类的自组织映射、用于线性或非线性时间序列预测的递归网络和适用于科学数据分析的其他类型的网络。 《神经网络在应用科学和工程中的应用从基本原理到复杂的模式识别》通过使用广泛的图示和多学科的内容以一种更容易理解的形式，填补了市场上神经网络用于多维科学数据的空白，并将神经网络与统计学联系了起来。 国际视野，科技前沿。 国际电气工程先进技术译丛，传播国际最新技术成果，搭建电气工程技术平台。 《神经网络在应用科学和工程中的应用从基本原理到复杂的模式识别》特点： ◆在多学科领域解释了神经网络； ◆为了易于理解，使用了大量图例来解释复杂数据概念； ◆深入研究了神经网络在线性和非线性预报、分类，聚类和预测方面的应用； ◆阐述了模型开发的所有阶段和结果的解释，包括数据预处理，数据维数约简，输入选择，模型开发和验证，模型不确定性评估以及对输入、误差和模型参数的灵敏度分析。

《神经网络与机器学习(英文版第3版)》的可读性非常强，作者举重若轻地对神经网络的基本模型和主要学习理论进行了深入探讨和分析，通过大量的试验报告、例题和习题来帮助读者更好地学习神经网络。神经网络是计算智能和机器学习的重要分支，在诸多领域都取得了很大的成功。在众多神经网络著作中，影响最为广泛的是SimonHaykin的《神经网络原理》(第4版更名为《神经网络与机器学习》）。在《神经网络与机器学习(英文版第3版)》中，作者结合近年来神经网络和机器学习的最新进展，从理论和实际应用出发，全面。系统地介绍了神经网络的基本模型、方法和技术，并将神经网络和机器学习有机地结合在一起。《神经网络与机器学习(英文版第3版)》不但注重对数学分析方法和理论的探讨，而且也非常关注神经网络在模式识别、信号处理以及控制系统等实际工程问题中的应用。 本版在前一版的基础上进行了广泛修订，提供了神经网络和机器学习这两个越来越重要的学科的最新分析。