20121005,微软十月份安全补丁提前通知

微软原生集成 AI Agent 的终端基础认识 在中,介绍了微软原生集成 AI Agent 的终端 Intelligent Terminal。项目地址:Intelligent Terminal 是微软推出的实验性终端产品,它是 Windows Terminal 的开源分支,通过 Agent Client Protocol (ACP) 把第三方 AI Coding Agent(Copilot、Claude、Codex、Gemini、OpenCode 等)以“一等公民”的方式直接嵌入终端界面。 阅读详情

大家好,我是 Richard Chen。

在此提前通知各位:微软计划于北京时间10月10日清晨发布7个安全补丁,共修复 Microsoft Word, Windows, Microsoft Office 和 SQL Server 中的20个安全漏洞。7个补丁的最高严重等级详见下图:

Bulletin IDMaximum Severity Rating and Vulnerability ImpactRestart RequirementAffected Software
Bulletin 1Critical
Remote Code Execution
May require restartMicrosoft Office,
Microsoft Server Software
Bulletin 2Important
Remote Code Execution
May require restartMicrosoft Office
Bulletin 3Important
Elevation of Privilege
May require restartMicrosoft Office,
Microsoft Server Software,
Microsoft Lync
Bulletin 4Important
Remote Code Execution
May require restartMicrosoft Office,
Microsoft Server Software
Bulletin 5Important
Elevation of Privilege
Requires restartMicrosoft Windows
Bulletin 6Important
Denial of Service
Requires restartMicrosoft Windows
Bulletin 7Important
Elevation of Privilege
May require restartMicrosoft SQL Server

按照受影响的操作系统分类如下:

Windows XP
Bulletin IdentifierBulletin 5Bulletin 6
Aggregate Severity RatingImportantNone
Windows XP Service Pack 3Windows XP Service Pack 3
(Important)
Not applicable
Windows XP Professional x64 Edition Service Pack 2Windows XP Professional x64 Edition Service Pack 2
(Important)
Not applicable
Windows Server 2003
Bulletin IdentifierBulletin 5Bulletin 6
Aggregate Severity RatingImportantNone
Windows Server 2003 Service Pack 2Windows Server 2003 Service Pack 2
(Important)
Not applicable
Windows Server 2003 x64 Edition Service Pack 2Windows Server 2003 x64 Edition Service Pack 2
(Important)
Not applicable
Windows Server 2003 with SP2 for Itanium-based SystemsWindows Server 2003 with SP2 for Itanium-based Systems
(Important)
Not applicable
Windows Vista
Bulletin IdentifierBulletin 5Bulletin 6
Aggregate Severity RatingImportantNone
Windows Vista Service Pack 2Windows Vista Service Pack 2
(Important)
Not applicable
Windows Vista x64 Edition Service Pack 2Windows Vista x64 Edition Service Pack 2
(Important)
Not applicable
Windows Server 2008
Bulletin IdentifierBulletin 5Bulletin 6
Aggregate Severity RatingImportantNone
Windows Server 2008 for 32-bit Systems Service Pack 2Windows Server 2008 for 32-bit Systems Service Pack 2
(Important)
Not applicable
Windows Server 2008 for x64-based Systems Service Pack 2Windows Server 2008 for x64-based Systems Service Pack 2
(Important)
Not applicable
Windows Server 2008 for Itanium-based Systems Service Pack 2Windows Server 2008 for Itanium-based Systems Service Pack 2
(Important)
Not applicable
Windows 7
Bulletin IdentifierBulletin 5Bulletin 6
Aggregate Severity RatingImportantImportant
Windows 7 for 32-bit SystemsWindows 7 for 32-bit Systems
(Important)
Windows 7 for 32-bit Systems
(Important)
Windows 7 for 32-bit Systems Service Pack 1Windows 7 for 32-bit Systems Service Pack 1
(Important)
Windows 7 for 32-bit Systems Service Pack 1
(Important)
Windows 7 for x64-based SystemsWindows 7 for x64-based Systems
(Important)
Windows 7 for x64-based Systems
(Important)
Windows 7 for x64-based Systems Service Pack 1Windows 7 for x64-based Systems Service Pack 1
(Important)
Windows 7 for x64-based Systems Service Pack 1
(Important)
Windows Server 2008 R2
Bulletin IdentifierBulletin 5Bulletin 6
Aggregate Severity RatingImportantImportant
Windows Server 2008 R2 for x64-based SystemsWindows Server 2008 R2 for x64-based Systems
(Important)
Windows Server 2008 R2 for x64-based Systems
(Important)
Windows Server 2008 R2 for x64-based Systems Service Pack 1Windows Server 2008 R2 for x64-based Systems Service Pack 1
(Important)
Windows Server 2008 R2 for x64-based Systems Service Pack 1
(Important)
Windows Server 2008 R2 for Itanium-based SystemsWindows Server 2008 R2 for Itanium-based Systems
(Important)
Windows Server 2008 R2 for Itanium-based Systems
(Important)
Windows Server 2008 R2 for Itanium-based Systems Service Pack 1Windows Server 2008 R2 for Itanium-based Systems Service Pack 1
(Important)
Windows Server 2008 R2 for Itanium-based Systems Service Pack 1
(Important)
Server Core installation option
Bulletin IdentifierBulletin 5Bulletin 6
Aggregate Severity RatingImportantImportant
Windows Server 2008 for 32-bit Systems Service Pack 2 (Server Core installation)Windows Server 2008 for 32-bit Systems Service Pack 2 (Server Core installation)
(Important)
Not applicable
Windows Server 2008 for x64-based Systems Service Pack 2 (Server Core installation)Windows Server 2008 for x64-based Systems Service Pack 2 (Server Core installation)
(Important)
Not applicable
Windows Server 2008 R2 for x64-based Systems (Server Core installation)Windows Server 2008 R2 for x64-based Systems (Server Core installation)
(Important)
Windows Server 2008 R2 for x64-based Systems (Server Core installation)
(Important)
Windows Server 2008 R2 for x64-based Systems Service Pack 1 (Server Core installation)Windows Server 2008 R2 for x64-based Systems Service Pack 1 (Server Core installation)
(Important)
Windows Server 2008 R2 for x64-based Systems Service Pack 1 (Server Core installation)
(Important)

微软 Office 补丁相关信息:

Microsoft Office Suites and Components
Bulletin IdentifierBulletin 1Bulletin 2Bulletin 3
Aggregate Severity RatingCriticalNoneNone
Microsoft Office 2003 Service Pack 3Microsoft Word 2003 Service Pack 3
(Important)
Not applicableNot applicable
Microsoft Office 2007 Service Pack 2Microsoft Word 2007 Service Pack 2
(Critical)
Not applicableNot applicable
Microsoft Office 2007 Service Pack 3Microsoft Word 2007 Service Pack 3
(Critical)
Not applicableNot applicable
Microsoft Office 2010 Service Pack 1 (32-bit editions)Microsoft Word 2010 Service Pack 1 (32-bit editions)
(Critical)
Not applicableNot applicable
Microsoft Office 2010 Service Pack 1 (64-bit editions)Microsoft Word 2010 Service Pack 1 (64-bit editions)
(Critical)
Not applicableNot applicable
Other Microsoft Office Software
Bulletin IdentifierBulletin 1Bulletin 2Bulletin 3
Aggregate Severity RatingImportantImportantImportant
Microsoft Word ViewerMicrosoft Word Viewer
(Important)
Not applicableNot applicable
Microsoft Office Compatibility Pack Service Pack 2Microsoft Office Compatibility Pack Service Pack 2
(Important)
Not applicableNot applicable
Microsoft Office Compatibility Pack Service Pack 3Microsoft Office Compatibility Pack Service Pack 3
(Important)
Not applicableNot applicable
Microsoft InfoPath 2007 Service Pack 2Not applicableNot applicableMicrosoft InfoPath 2007 Service Pack 2
(Important)
Microsoft InfoPath 2007 Service Pack 3Not applicableNot applicableMicrosoft InfoPath 2007 Service Pack 3
(Important)
Microsoft InfoPath 2010 Service Pack 1 (32-bit editions)Not applicableNot applicableMicrosoft InfoPath 2010 Service Pack 1 (32-bit editions)
(Important)
Microsoft InfoPath 2010 Service Pack 1 (64-bit editions)Not applicableNot applicableMicrosoft InfoPath 2010 Service Pack 1 (64-bit editions)
(Important)
Microsoft Works 9Not applicableMicrosoft Works 9
(Important)
Not applicable

Bulletin 1 的注释 : 本补丁影响多类软件。

Bulletin 3 的注释 : 本补丁影响多类软件。

微软服务器软件补丁相关信息:

Microsoft SharePoint Server
Bulletin IdentifierBulletin 1Bulletin 3Bulletin 4
Aggregate Severity RatingImportantImportantNone
Microsoft SharePoint Server 2007 Service Pack 2 (32-bit editions)Not applicableMicrosoft SharePoint Server 2007 Service Pack 2 (32-bit editions)
(Important)
Not applicable
Microsoft SharePoint Server 2007 Service Pack 3 (32-bit editions)Not applicableMicrosoft SharePoint Server 2007 Service Pack 3 (32-bit editions)
(Important)
Not applicable
Microsoft SharePoint Server 2007 Service Pack 2 (64-bit editions)Not applicableMicrosoft SharePoint Server 2007 Service Pack 2 (64-bit editions)
(Important)
Not applicable
Microsoft SharePoint Server 2007 Service Pack 3 (64-bit editions)Not applicableMicrosoft SharePoint Server 2007 Service Pack 3 (64-bit editions)
(Important)
Not applicable
Microsoft SharePoint Server 2010 Service Pack 1Microsoft SharePoint Server 2010 Service Pack 1
(Important)
Microsoft SharePoint Server 2010 Service Pack 1
(Important)
Not applicable
Microsoft FAST Search Server
Bulletin IdentifierBulletin 1Bulletin 3Bulletin 4
Aggregate Severity RatingNoneNoneImportant
Microsoft FAST Search Server 2010 for SharePointNot applicableNot applicableAdvanced Filter Pack
(Important)
Microsoft Groove Server
Bulletin IdentifierBulletin 1Bulletin 3Bulletin 4
Aggregate Severity RatingNoneImportantNone
Microsoft Groove Server 2010 Service Pack 1Not applicableMicrosoft Groove Server 2010 Service Pack 1
(Important)
Not applicable
Windows SharePoint Services and Microsoft SharePoint Foundation
Bulletin IdentifierBulletin 1Bulletin 3Bulletin 4
Aggregate Severity RatingNoneImportantNone
Microsoft Windows SharePoint Services 3.0 Service Pack 2 (32-bit version)Not applicableMicrosoft Windows SharePoint Services 3.0 Service Pack 2 (32-bit version)
(Important)
Not applicable
Microsoft Windows SharePoint Services 3.0 Service Pack 2 (64-bit version)Not applicableMicrosoft Windows SharePoint Services 3.0 Service Pack 2 (64-bit version)
(Important)
Not applicable
Microsoft SharePoint Foundation 2010 Service Pack 1Not applicableMicrosoft SharePoint Foundation 2010 Service Pack 1
(Important)
Not applicable
Microsoft Office Web Apps
Bulletin IdentifierBulletin 1Bulletin 3Bulletin 4
Aggregate Severity RatingImportantImportantNone
Microsoft Office Web Apps 2010 Service Pack 1Microsoft Office Web Apps 2010 Service Pack 1
(Important)
Microsoft Office Web Apps 2010 Service Pack 1
(Important)
Not applicable

Bulletin 1 的注释 : 本补丁影响多类软件。

Bulletin 3 的注释 : 本补丁影响多类软件。

微软通信平台和软件相关信息:

Microsoft Communicator
Bulletin IdentifierBulletin 3
Aggregate Severity RatingImportant
Microsoft Communicator 2007 R2Microsoft Communicator 2007 R2
(Important)
Microsoft Lync
Bulletin IdentifierBulletin 3
Aggregate Severity RatingImportant
Microsoft Lync 2010 (32-bit)Microsoft Lync 2010 (32-bit)
(Important)
Microsoft Lync 2010 (64-bit)Microsoft Lync 2010 (64-bit)
(Important)
Microsoft Lync 2010 AttendeeMicrosoft Lync 2010 Attendee
(Important)

Bulletin 3 的注释 : 本补丁影响多类软件。

微软SQL Server 相关信息:

以下为提前通知的文章全文(英文),请各位先行评估了解受影响的系统。

Microsoft Security Bulletin Advance Notification for Oct 2012:

http://technet.microsoft.com/en-us/security/bulletin/ms12-oct

Richard Chen

大中华区软件安全项目经理

DevCAI SimpleButton / CheckButton 按钮实现分析 分析对象:DevCAI 自研的与(对应第 2.3 节“选项、按钮与状态”中同名组件)(2286 行,单文件承载整个按钮家族)分析日期:2026-11-05技术栈:VS2022 / .NET Framework 4.8.1 / C# 8.0(库代码语法不高于 C# 7.3) 阅读详情

相关推荐

【AI智能体】Langchain 上下文记忆使用实战操作详解

【AI智能体】Langchain 上下文记忆使用实战操作详解

congge_study的博客 308

微软技术日报 2026-11-05:安装器变现链被拆开、OneLake 托管 MCP、Agent 365 进 Gartner 象限

把这几条放在一起看,共同点是把已经铺开的东西补上治理。安装器那条最直白。用户自己点开的程序,装完还会往回连,落地的东西藏在计划任务里、Program Files 里,用户不知道。微软给的检测思路也是往这个方向走的,看临时目录、看回调、看 SYSTEM 任务,而不是追域名。OneLake 那条是给 agent 补路标。以前每种 item 类型都要维护一份端点清单,现在让服务自己报出来。好处是 agent 不用猜,代价是判断责任转移到了客户端,技能要不要加载得自己决定,标签也不能当成授权。

海盗Sharp的博客 1421

应用程序向导

应用程序向导  何志丹            使用应用程序向导可以快速生成程序框架。选择菜单“File→New”命令,此时会弹出“New”对话框,下面对该对话框中的常用选项进行逐一说明。1.文件(Files)属性页 单击“Files”标签,打开“Files”属性页,如图1-3所示。图1-3  文件属性页文件类型虽然多,但常用的只有两种:C/C++ Header Fi

闻缺陷则喜何志丹 3698

微软技术日报 11-09:.NET 8 与 9 明日同时停止支持,注册表编辑器出了 WinUI 3 版

今天这四条消息的共同点是收口。.NET 8 和 9 的支持期到点,注册表编辑器的新界面只放在仓库里给人看,GitHub 把一类危险触发器按住默认,微软在安全报告里把防线从地址黑名单挪到行为。收口并不等于关门。.NET 10 还在支持期里,被按住的触发器可以用策略按文件放回来,实验室里的样例随时可能变成产品。区别在于这些动作要有人签字,不再是默认开着。如果今天只做一件事,去看一眼构建机上的 .NET 版本。明天之后,8 和 9 的漏洞补丁不会再来了。

海盗Sharp的博客 321

产业园区如何高效对接企业技术需求与高校科研成果?.docx

科易网基于AI模型+行业大数据,在技术转移、成果转化、技术经纪、知识产权、产业创新、科技招商等垂直领域的多样化应用场景,推出科技创新领域的AI+工具与数智化解决方案,推动科技创新与产业创新智能化发展。

100个Python代码大全:学完这篇从Python小白到架构师

Python是一种广泛使用的高级编程语言,以其简洁优雅、易于学习和强大的库而闻名。无论你是一位经验丰富的程序员还是初学者,都可以从这种语言的丰富资源中受益。在本篇博客中,我将分享100个实用的Python代码示例,旨在帮助你掌握Python编程,并在日常工作中提高效率。

penggerhe的博客 1万+

AI驱动数据处理的安全边界:行业现状、风险权衡与本地化实践路径

本文基于第三方数据,分析AI用于数据处理的安全议题。大模型应用加速普及,但伴随数据出境、影子AI、幻觉与成本失控等风险。行业共识指向“脚本复用”路径:将分析逻辑本地化固化,实现数据可控、结果可复现、成本可降。产品如数以轻舟提供本地处理、多模型接入与一次付费永久使用方案。建议企业从数据落点、可控性、可复现性与合规性四维度评估选型,推动AI在安全边界内释放价值。

数据矿工-数据化运营博客 291

Jetson Nano 【1】入手开箱、烧系统

类似于半开箱哈哈 怀着小期待的心情,屁颠屁颠跑到村口领快递,拿到了这个香喷喷的Jetson Nano 开发板 开箱: 由于太激动了忘记拍点啥,于是在最后留了张全家福 板子的结构: 我觉得这个有必要记录一下,于是去官方文档拿了张图来:官方文档传送 1=SD card 的卡槽,位于第二层板子的背部,也就是大风扇屁股位置 2=40针扩展接头(还没开用) 3=USB的供电口,根据大佬测试(某哩...

椰子奶糖的博客 4195

微软技术日报 11-08:Foundry 与 Entra 各挂一个 10.0 分漏洞,Foundry 接 fal 五款视频模型

这一周的动作可以分成两类。一类在收。Copilot 从电视上退场,搜索里的 AI 改成可选,云端漏洞只发通知不收补丁。共同点是承认铺得太开,代价比预期高。另一类在放。Foundry 一次摆上五个视频模型,容器版应用网关一个入口挂 64 个 MCP 服务器。共同点是把能力做成可组合的积木,让使用者按需挑。两个方向背后是同一个判断。把能力塞到每一个界面上,短期看覆盖率高,长期要解释为什么删不掉。新的做法是先给一个能关的开关,再谈覆盖。对每天写代码的人来说,今天有两样现成的东西可以试。

海盗Sharp的博客 374

别再用UE5默认材质:用Substance Painter一招提升写实感

今天,我就把这套从模型准备到引擎最终呈现的全流程技巧毫无保留地分享出来,这套方法不仅能让你的作品细节炸裂,更能让你理解次世代PBR(Physically Based Rendering)材质的底层逻辑。

海外旅居设计师,15年设计经验。 1325

手把手教你学PCIE(11.3)--pcie的认证和合规性测试

PCIe 设备的认证和合规性测试是确保设备正常工作和达到预期性能的关键步骤。通过使用 PCI-SIG 提供的测试工具和第三方测试工具,可以全面评估 PCIe 设备的物理层、链路层、事务层和互操作性,确保设备符合行业标准和规范。

MHD0815的博客 1360

微软技术周报 11-02~11-09:VS Code 混装 Claude 与 Codex、EF Core 11 静默改行为、.NET 11 倒计时

11 月 6 日,Power Platform 通过消息中心通知管理员,Copilot Studio、Power Automate 和 Power Apps 里所有跑在 Anthropic Claude、xAI Grok 以及旧版 GPT 上的 prompt,12 月 2 日自动迁移到 GPT-6 系列,档位价格不变,模型选择收窄到一条产品线。效果是这样的:让 Claude 评审一个 Copilot chat 写出的方案,或者把同样的任务交给 Codex,两边并排跑在同一个会话里。更要紧的是下周的时间线。

海盗Sharp的博客 831

Python对象的基本组成和内存示意图及引用的本质

在Python中,变量也称为:对象的引用(reference)。变量存储的就是对象的地址。变量不需要显式声明类型。根据变量引用的对象,Python解释器自动确定数据类型。一个内存块,拥有特定的值,支持特定类型的相关操作。变量位于:栈内存(压栈出栈等细节,后续再介绍)。⚠️Python是动态类型语言。变量通过地址引用了“对象”。

Calor13的博客 557

【测评】SenseNova6.8Flash模型正式发布,支持动态 PPT 及 3D 设计

SenseNova 6.8 Flash将设计生成拓展至多种创作形态,支持Web、SVG、Three.js、Blender、PPT等形式,满足多样化的AI Native Design需求,并支持动态PPT、动态Poster与Agentic 3D设计。同时,模型可接入Figma等专业工作流,帮助用户将AI设计能力融入现有的专业设计流程。此外,模型还支持三维场景与交互内容创作,可综合处理空间布局、材质与光照、镜头控制及动画编排等因素,构建可在网页中实时渲染的三维内容,拓展从立体视觉展示到动态交互的表达方式。

Enweitech Software Works 337

【干货分享】高通 SA8295P PCB 设计规范指南

高通SA8295P芯片是第四代骁龙汽车数字座舱平台的核心组件,采用5nm工艺,集成高性能CPU/GPU和30TOPS算力的NPU。文章详细介绍了该芯片各接口的layout设计规范,包括USB3.1/2.0、eDP/DP、MIPI CSI/DSI、PCIe Gen3等接口的阻抗、损耗、延时等关键指标要求,为新能源汽车电子设计提供重要参考。

一只豌豆象的博客 265

RK3588网卡出现丢包或者带宽不足的问题验证排查

注意除了有最大的TCP数据接收缓冲,还有/proc/sys/net/core/wmem_max — 最大的TCP数据发送缓冲可以调整,这里暂时没用到,针对具体需要可适当调整。比如把ethx接口中断从默认的cpu0迁移到其它cpu,例如将eth0中断234移到cpu7 (10是16进制, cpu0 =3588的cpu有8个核,4个A55(小核)+4个A76(大核),可以找到对应中断 将中断从默认cpu0核切到cpu7。tcp_rmem:单位是字节,接收缓存区大小,缓存从对端接收的数据,后续会被应用程序读取。

Android/Linux驱动开发传帮带学习,让驱动入门门槛更低! 2029

微软输入法标点乱跳解决方案

微软拼音输入法常因“中文输入时使用英文标点”开关被意外开启,导致标点自动切换为英文且快捷键失效。本文提供一键批处理脚本,通过修改注册表(HKCU\Software\Microsoft\InputMethod\Settings\CHS)强制锁定中文标点,覆盖多个相关键值,确保设置持久生效。相比手动操作,该方案可抵御更新、同步等导致的配置重置,提升稳定性。执行后需重启输入法或系统以使更改生效,建议提前备份注册表。

谙弆悕博士的技术博客! 355

【HALCON工业视觉应用探索】02. C# + Halcon铸件毛刺机器人打磨引导:点云处理从实验室到工业现场的全流程落地

摘要:铸件毛刺的自动化打磨是机械制造领域的重要难题,其随机性(高度0.1-3mm)和复杂工况导致人工打磨效率低(3分钟/件)、工伤率高。本文基于Halcon 24.11与.NET 6,构建了一套点云处理方案:通过线激光扫描仪获取三维点云,经预处理、毛刺检测、路径规划三大模块,引导机器人完成打磨。虚拟实验测试显示,对高度>1mm的毛刺检出率达100%,定位误差±0.08mm,但工业现场受油污、粉尘影响,性能降至检出率68-75%、处理时间>45秒。

专注于人工智能、软件开发、工控自动化、工厂数字化及智能化等领域,希望和大家共同进步! 1128

微软 BioEmu2 实测:AI 设计的抗体,能否通过动态模拟提前筛选?

微软 BioEmu2 实测:AI 设计的抗体,能否通过动态模拟提前筛选?**2026 年 10 月 26 日,展示了一个 AI 抗体候选的 10 纳秒传播动画。**但这个案例留下的最大问题,仍是实验室门前那道选择题:一批看起来合理的设计,究竟该把哪几个送去测试?Ford 正在参与 Adaptyv 与 Anthropic 的蛋白设计竞赛,也在开发抗体语言模型。**但他在原文中坦言,生成候选容易,挑出值得测试的候选依然困难。**本文解读他的公开实测,并对照微软文档核查证据边界,未独立运行模型。

weixin_71777199的博客 200

WINCC7.3破解文件

WINCC7.3硬件够破解文件,有需要的下载!WINCC7.3硬件够破解文件,有需要的下载!

SIMATIC S7-1200, S7-1500 PID 控制[手册]

SIMATIC S7-1200, S7-1500 PID 控制[手册]

视频美颜sdk的API设计会影响开发效率吗?

视频美颜SDK不仅需要具备优秀的图像处理能力,API接口设计同样会影响软件开发效率。本文从接口调用、参数配置、版本兼容、开发文档以及业务扩展等多个角度,分析API设计对视频美颜SDK集成和后期维护的实际影响,为直播APP、视频社交平台等项目的SDK选型与技术开发提供参考。

wanyuekeji的博客 218

深度学习工具包 DeepLearningToolbox

深度学习工具包Deprecation notice.-----This toolbox is outdated and no longer maintained.There are much better tools available for deep learning than this toolbox, e.g. [Theano](http://deeplearning.net/software/theano/), [torch](http://torch.ch/) or [tensorflow](http://www.tensorflow.org/)I would suggest you use one of the tools mentioned above rather than use this toolbox.Best, Rasmus.DeepLearnToolbox================A Matlab toolbox for Deep Learning.Deep Learning is a new subfield of machine learning that focuses on learning deep hierarchical models of data.It is inspired by the human brain's apparent deep (layered, hierarchical) architecture.A good overview of the theory of Deep Learning theory is[Learning Deep Architectures for AI](http://www.iro.umontreal.ca/~bengioy/papers/ftml_book.pdf)For a more informal introduction, see the following videos by Geoffrey Hinton and Andrew Ng.* [The Next Generation of Neural Networks](http://www.youtube.com/watch?v=AyzOUbkUf3M) (Hinton, 2007)* [Recent Developments in Deep Learning](http://www.youtube.com/watch?v=VdIURAu1-aU) (Hinton, 2010)* [Unsupervised Feature Learning and Deep Learning](http://www.youtube.com/watch?v=ZmNOAtZIgIk) (Ng, 2011)If you use this toolbox in your research please cite [Prediction as a candidate for learning deep hierarchical models of data](http://www2.imm.dtu.dk/pubdb/views/publication_details.php?id=6284)```@MASTERSTHESIS\{IMM2012-06284, author = "R. B. Palm", title = "Prediction as a candidate for learning deep hierarchical models of data", year = "2012",}```Contact: rasmusbergpalm at gmail dot comDirectories included in the toolbox-----------------------------------`NN/` - A library for Feedforward Backpropagation Neural Networks`CNN/` - A library for Convolutional Neural Networks`DBN/` - A library for Deep Belief Networks`SAE/` - A library for Stacked Auto-Encoders`CAE/` - A library for Convolutional Auto-Encoders`util/` - Utility functions used by the libraries`data/` - Data used by the examples`tests/` - unit tests to verify toolbox is workingFor references on each library check REFS.mdSetup-----1. Download.2. addpath(genpath('DeepLearnToolbox'));Example: Deep Belief Network---------------------```matlabfunction test_example_DBNload mnist_uint8;train_x = double(train_x) / 255;test_x = double(test_x) / 255;train_y = double(train_y);test_y = double(test_y);%% ex1 train a 100 hidden unit RBM and visualize its weightsrand('state',0)dbn.sizes = [100];opts.numepochs = 1;opts.batchsize = 100;opts.momentum = 0;opts.alpha = 1;dbn = dbnsetup(dbn, train_x, opts);dbn = dbntrain(dbn, train_x, opts);figure; visualize(dbn.rbm{1}.W'); % Visualize the RBM weights%% ex2 train a 100-100 hidden unit DBN and use its weights to initialize a NNrand('state',0)%train dbndbn.sizes = [100 100];opts.numepochs = 1;opts.batchsize = 100;opts.momentum = 0;opts.alpha = 1;dbn = dbnsetup(dbn, train_x, opts);dbn = dbntrain(dbn, train_x, opts);%unfold dbn to nnnn = dbnunfoldtonn(dbn, 10);nn.activation_function = 'sigm';%train nnopts.numepochs = 1;opts.batchsize = 100;nn = nntrain(nn, train_x, train_y, opts);[er, bad] = nntest(nn, test_x, test_y);assert(er < 0.10, 'Too big error');```Example: Stacked Auto-Encoders---------------------```matlabfunction test_example_SAEload mnist_uint8;train_x = double(train_x)/255;test_x = double(test_x)/255;train_y = double(train_y);test_y = double(test_y);%% ex1 train a 100 hidden unit SDAE and use it to initialize a FFNN% Setup and train a stacked denoising autoencoder (SDAE)rand('state',0)sae = saesetup([784 100]);sae.ae{1}.activation_function = 'sigm';sae.ae{1}.learningRate = 1;sae.ae{1}.inputZeroMaskedFraction = 0.5;opts.numepochs = 1;opts.batchsize = 100;sae = saetrain(sae, train_x, opts);visualize(sae.ae{1}.W{1}(:,2:end)')% Use the SDAE to initialize a FFNNnn = nnsetup([784 100 10]);nn.activation_function = 'sigm';nn.learningRate = 1;nn.W{1} = sae.ae{1}.W{1};% Train the FFNNopts.numepochs = 1;opts.batchsize = 100;nn = nntrain(nn, train_x, train_y, opts);[er, bad] = nntest(nn, test_x, test_y);assert(er < 0.16, 'Too big error');```Example: Convolutional Neural Nets---------------------```matlabfunction test_example_CNNload mnist_uint8;train_x = double(reshape(train_x',28,28,60000))/255;test_x = double(reshape(test_x',28,28,10000))/255;train_y = double(train_y');test_y = double(test_y');%% ex1 Train a 6c-2s-12c-2s Convolutional neural network %will run 1 epoch in about 200 second and get around 11% error. %With 100 epochs you'll get around 1.2% errorrand('state',0)cnn.layers = { struct('type', 'i') %input layer struct('type', 'c', 'outputmaps', 6, 'kernelsize', 5) %convolution layer struct('type', 's', 'scale', 2) %sub sampling layer struct('type', 'c', 'outputmaps', 12, 'kernelsize', 5) %convolution layer struct('type', 's', 'scale', 2) %subsampling layer};cnn = cnnsetup(cnn, train_x, train_y);opts.alpha = 1;opts.batchsize = 50;opts.numepochs = 1;cnn = cnntrain(cnn, train_x, train_y, opts);[er, bad] = cnntest(cnn, test_x, test_y);%plot mean squared errorfigure; plot(cnn.rL);assert(er<0.12, 'Too big error');```Example: Neural Networks---------------------```matlabfunction test_example_NNload mnist_uint8;train_x = double(train_x) / 255;test_x = double(test_x) / 255;train_y = double(train_y);test_y = double(test_y);% normalize[train_x, mu, sigma] = zscore(train_x);test_x = normalize(test_x, mu, sigma);%% ex1 vanilla neural netrand('state',0)nn = nnsetup([784 100 10]);opts.numepochs = 1; % Number of full sweeps through dataopts.batchsize = 100; % Take a mean gradient step over this many samples[nn, L] = nntrain(nn, train_x, train_y, opts);[er, bad] = nntest(nn, test_x, test_y);assert(er < 0.08, 'Too big error');%% ex2 neural net with L2 weight decayrand('state',0)nn = nnsetup([784 100 10]);nn.weightPenaltyL2 = 1e-4; % L2 weight decayopts.numepochs = 1; % Number of full sweeps through dataopts.batchsize = 100; % Take a mean gradient step over this many samplesnn = nntrain(nn, train_x, train_y, opts);[er, bad] = nntest(nn, test_x, test_y);assert(er < 0.1, 'Too big error');%% ex3 neural net with dropoutrand('state',0)nn = nnsetup([784 100 10]);nn.dropoutFraction = 0.5; % Dropout fraction opts.numepochs = 1; % Number of full sweeps through dataopts.batchsize = 100; % Take a mean gradient step over this many samplesnn = nntrain(nn, train_x, train_y, opts);[er, bad] = nntest(nn, test_x, test_y);assert(er < 0.1, 'Too big error');%% ex4 neural net with sigmoid activation functionrand('state',0)nn = nnsetup([784 100 10]);nn.activation_function = 'sigm'; % Sigmoid activation functionnn.learningRate = 1; % Sigm require a lower learning rateopts.numepochs = 1; % Number of full sweeps through dataopts.batchsize = 100; % Take a mean gradient step over this many samplesnn = nntrain(nn, train_x, train_y, opts);[er, bad] = nntest(nn, test_x, test_y);assert(er < 0.1, 'Too big error');%% ex5 plotting functionalityrand('state',0)nn = nnsetup([784 20 10]);opts.numepochs = 5; % Number of full sweeps through datann.output = 'softmax'; % use softmax outputopts.batchsize = 1000; % Take a mean gradient step over this many samplesopts.plot = 1; % enable plottingnn = nntrain(nn, train_x, train_y, opts);[er, bad] = nntest(nn, test_x, test_y);assert(er < 0.1, 'Too big error');%% ex6 neural net with sigmoid activation and plotting of validation and training error% split training data into training and validation datavx = train_x(1:10000,:);tx = train_x(10001:end,:);vy = train_y(1:10000,:);ty = train_y(10001:end,:);rand('state',0)nn = nnsetup([784 20 10]); nn.output = 'softmax'; % use softmax outputopts.numepochs = 5; % Number of full sweeps through dataopts.batchsize = 1000; % Take a mean gradient step over this many samplesopts.plot = 1; % enable plottingnn = nntrain(nn, tx, ty, opts, vx, vy); % nntrain takes validation set as last two arguments (optionally)[er, bad] = nntest(nn, test_x, test_y);assert(er < 0.1, 'Too big error');```[![Bitdeli Badge](https://d2weczhvl823v0.cloudfront.net/rasmusbergpalm/deeplearntoolbox/trend.png)](https://bitdeli.com/free "Bitdeli Badge")

微软技术日报 2026-11-04:8GB 专属优化、38 万漏洞利用统计、NuGet 包带技能

11 月的第一周,微软把注意力放回了最不显眼的那批机器。11 月 3 日,Windows 官方博客发出一篇由 Pavan Davuluri 署名的文章。他是微软执行副总裁,负责 Windows 与设备。文章的题目直白得少见,讲的是怎么把 8GB 内存的 Windows 11 电脑拉到及格线以上。同一天,.NET 博客在讲另一件事,怎么把 agent skills 塞进 NuGet 包。11 月 2 日,微软安全研究发了一份跨 27 年的漏洞利用统计。

海盗Sharp的博客 466

文档:亲测可用解决vnc灰屏,黑屏,鼠标大黑叉情况

主要解决vnc-viewer灰屏,黑屏,鼠标大黑叉等现象,详细见文档。

微软技术日报 2026-11-06:本地服务器补丁将自动安装、画图 8.1 分漏洞修复卡在商店、DACS 开源

今天这四条,能拼出一条共同的线:默认值在往自动那一侧挪。补丁从等人装改成到期自动装,紧急缓解从需要开启改成默认开启。智能体的权限从长期持有改成按任务发、按风险收,普通用户的提权从整块管理员切成改一个网络设置这样的单点。这条线有同一个前提,策略得先写对。自动装补丁的前提是你能接受某天早上服务器自己重启;按任务发权限的前提是任务边界描述得足够清楚。默认值变聪明,不等于判断变简单。还有一条更细的教训来自画图那个漏洞。修复版本发布在商店里,公告写着需要客户操作,但商店里那几天没有可装的东西。

海盗Sharp的博客 618

微软技术日报 2026-11-07:Copilot 本机跑 27B 模型、Front Door 默认后量子 TLS、Beta 转 26H2

四条消息挤在同一天,方向各不相同:Foundry Local 把 27B 的编码模型压进了 24GB 显存并接进 GitHub Copilot;Azure Front Door 把入站 TLS 1.3 的混合后量子密钥交换设成了默认;Windows Insider 一次放出五个构建,Beta 通道正式转到 26H2;Power Platform 通知管理员,12 月 2 日提示模型换血,Claude 与 Grok 一起退出。四条放在一起看,有两条在改默认值,两条在改节奏。

海盗Sharp的博客 458

Gemini 语音混音入门:在 AI Studio 用文字调整音色和口音

说明在 Google AI Studio 基于已有音色用自然语言做 Voice remixing:选择基础音色、输入相对调整、试听预览并保存为独立 voice ID,再用混音后的声音合成旁白。

死磕音视频 173

微软技术日报 11-10:.NET 11 今晚发布,.NET 8 与 9 同日停更,Copilot 连接器开写入

今天微软技术侧的重心全压在一个日期上。11 月 10 日既是 .NET Conf 2026 开幕、.NET 11 正式发布的日子,也是 .NET 8 与 .NET 9 终止支持的日子。同一个日期上,一条线开始,两条线结束。除了这条主线,还有三件事在这一两天落地。微软安全博客公布了一轮用智能体扫 Rust 标准库的研究,Copilot 连接器的写入能力转为正式可用,Windows 11 的统一应用更新页又在预览版里冒了头。

海盗Sharp的博客 813
上一篇: 微软安全新闻聚焦-双周刊第二十一期
下一篇: 20121010,微软10月10日发布7个安全补丁
MSSecurity
博客等级 码龄17年 613粉丝 300原创
评论
成就一亿技术人!
拼手气红包6.0元
还能输入1000个字符
 
 条评论被折叠 查看
添加红包

请填写红包祝福语或标题

个

红包个数最小为10个

元

红包金额最低5元

当前余额3.43元 前往充值 >
需支付:10.00元
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付元
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明:

1.余额是钱包充值的虚拟货币,按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载,可以购买VIP、付费专栏及课程。

余额充值