Swin变形金刚-PyTorch 体系结构的实现。 本文介绍了一种称为Swin Transformer的新型视觉变形金刚，它可以用作计算机视觉的通用骨干。 在两个领域之间的差异，例如视觉实体规模的巨大差异以及与文字中的单词相比，图像中像素的高分辨率，带来了使Transformer从语言适应视觉方面的挑战。 为了解决这些差异，我们提出了一个分层的Transformer，其表示是通过移动窗口来计算的。 通过将自注意计算限制为不重叠的局部窗口，同时允许跨窗口连接，移位的加窗方案带来了更高的效率。 这种分层体系结构具有在各种规模上建模的灵活性，并且相对于图像大小具有线性计算复杂性。 Swin Transformer的这些品质使其可与多种视觉任务兼容，包括图像分类（ImageNet-1K的准确度为86.4 top-1）和密集的预测任务，例如目标检测（COCO测试中为58.7框式AP和51.1遮罩式

人工智能导论（Introduction to Artificial Intelligence），722页ppt，核心内容包括：“人工智能：一种现代的方法”【第三版，作者：罗素 (Stuart J.Russell) / 诺维格 (Peter Norvig)】详细的讲了人工智能领域，并教会从头开始编写、调试和运行（某些）AI算法。

CS188.1x-人工智能（edX） 这些是 02.2015 edX/BerkleyX 课程的作业，可在此处找到： ://www.edx.org/course/artificial-intelligence-uc-berkeleyx-cs188-1x-0 项目一 深度优先搜索 (DFS) 广度优先搜索 (BFS) 统一成本搜索 (UCS) 一星搜索 (A*) 各种启发式 文件编辑： search.py searchAgents.py 项目二 反射剂的评价函数 具有多个对手的极小极大 Alpha Beta 修剪 平均期望值 状态（而不是动作）的评估函数 文件编辑： multiAgents.py 项目3 价值迭代离线规划代理 策略计算和参数 Q-学习 Epsilon 贪婪（q-learning） 近似 q-learning 和状态抽象 具有特征和权重的近似 q-learni

PROLOG PROGRAMMING FOR ARTIFICIAL INTELLIGENCE - lvan Bratko.pdf PROLOG PROGRAMMING FOR ARTIFICIAL INTELLIGENCE - lvan Bratko.pdf

Swarm Intelligence- From Natural to Artificial Systems

2020年4月22日，兰德发布了题为《Military Applications of Artificial Intelligence—Ethical Concerns in an Uncertain World》的报告。报告中介绍了美国、俄罗斯、中国人工智能在军事领域应用情况。

Jeff Hawkins - On Intelligence

The_Future_of_Cyber_Intelligence_How_Actionable_Deep_Learning_is_Changing_the_Face_of_Security AI 数据分析 移动安全 渗透测试 渗透测试

TinyML实现可识别自来水龙头的声音，一旦听到声音，便会触发蜂鸣器+ LED计时器。

Open Source Intelligence Methods and Tools focuses on building a deep understanding of how to exploit open source intelligence (OSINT) techniques, methods, and tools to acquire information from publicly available online sources to support intelligence analysis. The harvested data can be used in different scenarios such as financial, crime,and terrorism investigations as well as in more regular tasks such as analyzing business competitors, running background checks, and acquiring intelligence about individuals and other entities. This book will also improve your skills in acquiring information online from the surface web, the deep web, and the darknet. Many estimates show that 90 percent of useful information acquired by intelligence services comes from public sources (in other words, OSINT sources). Social media sites open up numerous opportunities for investigations because of the vast amount of useful information located in one place. For example, you can get a great deal of personal information about any person worldwide by just checking their Facebook page. This book will show you how to conduct advanced social media investigations to access content believed to be private, use advanced search engines queries to return accurate results, search historical deleted versions of websites, track individuals online using public record databases and people-searching tools, locate information buried in the deep web, access and navigate the dark web, collect intelligence from the dark web, view multiple historic satellite images and street views of any location, search geolocation information within popular social media sites, and more. In short, you will learn how to use a plethora of techniques, tools, and free online services to gather intelligence about any target online. OSINT-gathering activities should be conducted secretly to avoid revealing the searcher’s identity. Therefore, this book will teach you how to conceal your digital identity and become anonymous online