《人工智能基础与实践》
上海交通大学出版社印刷审核稿。主编：周兵、廖银萍、冯霞；14章在线教材，并配套首批5章“一章一项目”教学资源样板。
Higher-vocational edition in press review with Shanghai Jiao Tong University Press. Fourteen online chapters plus five pilot chapter resource packs.
职业教育人工智能课程与项目实践资源VOCATIONAL AI CURRICULUM & PROJECT RESOURCES
AI Education
Resource Portal
从通识教材到专业课程与物理AIFrom AI Literacy to Professional Curriculum and Physical AI
汇集正在出版和持续开发的人工智能教材、数字实验、教学资源与项目实践，形成“基础素养—工程方法—专业核心—Physical AI”的可组合课程体系。
A growing collection of AI textbooks, digital labs, teaching resources and project practice, organized into a composable pathway from foundational literacy and engineering methods to professional core learning and Physical AI.
01
AI基础与素养AI Literacy模型、数据、语言、多模态与责任边界Models, data, language, multimodality and responsible use
→
02
AI工程实践AI Engineering工具调用、智能体、工作流与证据化实践Tool use, agents, workflows and evidence-based practice
→
03
物理AI项目Physical AI Projects感知、边端计算、模型、反馈、测试与迭代Sensing, edge computing, models, feedback, testing and iteration
✓
核心原则：Core principle: AI参与学习和执行；关键判断、复核与最终责任由人承担。AI supports learning and execution; critical judgment, verification and final responsibility remain human.
5本教材5 Textbooks通识、基础与专业核心Literacy, foundations and core study
64章在线内容64 Online Chapters稳定HTML阅读路径Stable HTML reading routes
5章资源样板5 Resource PilotsNotebook、脚本与项目页Notebooks, scripts and project pages
双地址兼容URL Compatibility原教材网址继续有效Legacy book URLs remain live
持续更新Living Resources状态透明、分批发布Transparent, staged releases
教材与在线资源矩阵Textbook and Online Resource Collection
每本教材沿用原有稳定slug，并在本站建立长期入口。已出版信息与开发稿状态分别标注；所有未正式出版的在线版均设独立版权与版本页。
Each title keeps its stable slug and gains a long-term portal entry. Published metadata and manuscript status are distinguished, with a dedicated book-information page for every online edition.
《人工智能基础与实践（中职版）》
面向中等职业教育的14章项目化学习版本。主编：周兵；编写、审读与试教中，尚未正式出版。
A 14-chapter project-based secondary-vocational edition under writing, review and teaching trials.
《人工智能通识：理解、使用与判断》
12章全书初稿，面向不同专业学生。主编：周兵；拟对接国家开放大学出版社，尚未正式出版。
A 12-chapter AI literacy manuscript for learners across disciplines, currently under editorial review.
《人工智能工作方法与实践》
12章第一版完整稿，围绕工作任务、证据、验收与交付。主编：周兵；拟对接国家开放大学出版社，试教与外部证据补充中。
A 12-chapter complete first manuscript focused on work methods, evidence, acceptance and delivery.
《机器学习工程实践》
12章教学与出版审校版，涵盖机器学习、视觉、语言、边缘计算、机器人、智能体与系统工程。主编：周兵；书名暂定，出版社待定。
A 12-chapter professional-core review edition spanning machine learning, vision, language, edge AI, robotics, agents and systems engineering.
一套资源，三层进阶One resource, three progressive layers
学习内容由AI基础概念逐步进入工程实践和项目作品。三个层级可单独选用，也可连续组织为完整课程。
The learning path progresses from AI foundations to engineering practice and project artifacts. Each layer can be adopted independently or delivered as a continuous curriculum.
01 · AI基础与素养01 · AI Literacy
理解数据、模型、提示词、语义表示、RAG、多模态和AI幻觉，建立人机协同与安全边界意识。
Understand data, models, prompting, semantic representations, RAG, multimodality and hallucinations while building responsible human-AI collaboration habits.
02 · AI工程实践02 · AI Engineering
在可运行工作区中学习工具调用、沙盒、智能体、多步编排与AI辅助工作流，并完成测试、记录和复盘。
Use runnable workspaces to learn tool use, sandboxing, agents, multi-step orchestration and AI-assisted workflows, with testing, records and review built into the process.
03 · 物理AI项目03 · Physical AI Projects
把摄像头、传感器、边端计算与模型连接起来，形成感知—判断—反馈—日志—人工复核的最小闭环。
Connect cameras, sensors, edge computing and models into a minimum closed loop of sensing, decision, feedback, logging and human verification.
深圳实践，全球可迁移Shenzhen-rooted, globally adaptable
深圳在人工智能、智能终端、传感器、机器人和先进制造等领域具有完整的产业链和丰富的工程场景。课程将器件选择、端侧算力、通信、终端集成、测试与快速迭代等实际环节转化为教学任务，并将模型、硬件和平台设计为可替换模块，便于不同地区和院校按本地条件实施。
Shenzhen provides a rich engineering context across AI, intelligent terminals, sensors, robotics and advanced manufacturing. The curriculum turns component selection, edge compute, connectivity, terminal integration, testing and rapid iteration into learning tasks, with replaceable models, hardware and platforms for local adaptation.
产业场景进入项目教学Industry Context in Project Learning
学生在项目中理解模型如何进入终端产品，并分析功耗、算力、传感器、网络、成本和测试条件对系统设计的影响。
Learners examine how models are integrated into terminal products and how power, compute, sensors, networks, cost and testing conditions shape system design.
感知Sensing摄像头 · 麦克风 · 传感器Camera · Audio · Sensors
边端计算Edge ComputeMCU · NPU · 本地推理MCU · NPU · Local inference
模型Models识别 · 多模态 · AgentRecognition · Multimodal · Agent
通信Connectivity本地 · 网络 · 云边协同Local · Network · Cloud-edge
终端Terminals可穿戴 · 智能设备 · 机器人Wearables · Devices · Robotics
测试与迭代Test & Iterate验证 · 日志 · 复核 · 改进Validate · Log · Review · Improve
《人工智能基础与实践》AI Foundations and Practice
教材共14章，从AI时代的人机协同出发，依次涵盖数据、语言、提示词、RAG、多模态、工具调用、智能体、AI工作流和物理AI综合实践。
The textbook contains 14 chapters covering human-AI collaboration, data, language, prompting, RAG, multimodality, tool use, agents, AI workflows and Physical AI practice.
Shanghai Jiao Tong University Press
人工智能
基础与实践AI Foundations
& Practice
基础与实践AI Foundations
& Practice
高等职业教育 · 14章 · ISBN 978-7-313-35166-1
主编：周兵、廖银萍、冯霞Higher vocational education · 14 chapters · ISBN 978-7-313-35166-1
Editors-in-chief: Zhou Bing, Liao Yinping, Feng Xia
主编：周兵、廖银萍、冯霞Higher vocational education · 14 chapters · ISBN 978-7-313-35166-1
Editors-in-chief: Zhou Bing, Liao Yinping, Feng Xia
模块A · 人机协同与学习环境Module A · Human-AI Collaboration
第1—3章：职业能力、人机协同、行动循环、大模型学习环境。Ch. 1–3: professional capability, collaboration, action loops and AI learning environments.
模块B · 数据、语言与生成机制Module B · Data, Language & Generation
第4—7章：模型与预测、语义表示、提示词工程、上下文与幻觉。Ch. 4–7: models and prediction, semantic representations, prompting, context and hallucination.
模块C · 检索、多模态与可控执行Module C · Retrieval, Multimodality & Controlled Action
第8—11章：RAG、视觉理解、工具调用、沙盒、智能体与多步编排。Ch. 8–11: RAG, visual understanding, tool use, sandboxing, agents and orchestration.
模块D · 工作流与物理AI作品Module D · Workflows & Physical AI
第12—14章：AI辅助工作流、本地智能体、摄像头识别与智能信号卡作品。Ch. 12–14: AI-assisted workflows, local agents, camera recognition and a signal-card project.
AI协同学习与实践Human-AI Collaborative Learning and Practice
资源采用工程化的人机协同方式。AI用于生成方案、解释、代码建议和工具调用提议，学习者负责观察真实结果、执行测试、人工复核并保存证据。
The resource uses an engineering-oriented collaboration model. AI supports planning, explanation, code suggestions and tool-call proposals, while learners observe real outcomes, test, review and preserve evidence.
01理解问题Understand
明确目标与场景Define goal & context
02定义任务Specify
资料、约束与验收Inputs, constraints, acceptance
03AI协作Collaborate
计划、生成、调用工具Plan, generate, use tools
04运行实验Run
得到真实输出Observe real output
05测试Test
对照、异常与边界Compare, stress, challenge
06人工复核Review
事实、风险与责任Facts, risks, responsibility
07证据与交付Evidence
日志、版本与作品Logs, versions, artifacts
五类可组合资源Five composable resource types
院校可以整门采用，也可以按模块和项目组合。资源按“已可使用、教学试点、开发中”标注当前状态，便于了解可用范围和后续建设计划。
Institutions may adopt the full course or combine modules and projects. Resource status is marked as Available, Pilot or In Development to show current availability and planned extensions.
Available
课程与教材Curriculum & Textbook
在线教材、章节目标、学习任务与综合实践。Online textbook, learning outcomes, tasks and integrated practice.
Available
数字实验Digital Labs
Notebook、示例数据、脚本、实验记录模板。Notebooks, sample data, scripts and experiment records.
Pilot
AI协同工作区AI Collaboration Workspace
VS Code、模型入口、Git工作流与安全规则。VS Code, model access, Git workflows and safety rules.
Pilot
项目实践包Project Packs
任务书、需求、测试集、失败样例与证据包。Briefs, requirements, tests, failure cases and evidence packs.
In Development
教师实施资源Faculty Resources
实施指南、课堂组织、评价与专业进阶模块。Implementation guides, classroom organization, assessment and advanced modules.
代表项目：智能信号卡识别Flagship project: Smart Signal Card Recognition
用低风险、可演示的作品把物理AI概念落地：摄像头采集真实环境输入，本地程序完成识别与反馈，同时记录日志并保留人工复核点。
A low-risk, demonstrable project turns Physical AI into a concrete learning artifact: a camera captures real-world input, local software performs recognition and feedback, and logs preserve human verification points.
智能信号卡项目实践流程Smart Signal Card Project Workflow
项目采用较简单的模型，重点训练完整工程闭环以及学生对AI输出边界的判断。
The project uses a relatively simple model and focuses on the complete engineering loop and learners' judgment of AI output boundaries.
1摄像头输入Camera input
2识别与状态判断Recognition & state decision
3可见反馈Visible feedback
4运行日志Runtime log
5测试矩阵Test matrix
6人工复核与版本迭代Human review & iteration
学生最终交付Student Deliverables
- 可运行程序或项目文件Runnable program or project files
- 测试输入与结果记录Test inputs and results
- AI参与过程与人工修改说明AI involvement and human modification record
- 异常案例与失败复盘Failure cases and retrospective
- 隐私、安全与责任边界检查Privacy, safety and responsibility checks
课程采用、教师培训与联合开发Course Adoption, Faculty Development and Co-creation
资源采用模块化组织。合作学校可根据本地条件替换品牌、设备和模型服务，同时沿用课程结构、工程方法、项目任务与评价证据。
The resource is organized in modules. Partner institutions can adapt brands, devices and model services to local conditions while retaining the curriculum structure, engineering methods, project tasks and assessment evidence.
课程采用Course Adoption
整门课程或选取章节、实验和项目模块。Adopt the full course or selected chapters, labs and projects.
教师培训Faculty Training
围绕AI协同教学、工程工作流和项目实施开展培训。Training in human-AI teaching, engineering workflows and project delivery.
资源本地化Localization
替换语言、行业案例、模型服务和硬件模块。Adapt language, domain cases, model services and hardware modules.
联合项目Joint Development
共同开发AI Engineering与Physical AI实践项目。Co-develop AI Engineering and Physical AI learning projects.
资源建设状态Resource Status
Available表示当前可阅读或使用；Pilot表示已通过自动验收、公开试用但仍待教师审阅和学生试学；In Development表示正在开发的教材与新增项目。
Available indicates resources ready to read or use; Pilot indicates public samples that passed automated checks but still await faculty review and learner trials; In Development covers textbooks and projects under active development.
Available
当前可使用Ready now
- 《人工智能基础与实践》14章内容14-chapter AI Foundations & Practice textbook
- 在线HTML阅读Online HTML reading
- 核心实验与综合实践框架Core lab and integrated-practice framework
Pilot Samples
公开试用样板Public pilot samples
- 高职版第3、4、8、13、14章资源Higher-edition resources for Chapters 3, 4, 8, 13 and 14
- Notebook、项目页与离线回归Notebooks, project pages and offline regression checks
- 待教师审阅与学生试学Faculty review and learner trials pending
In Development
持续扩展Next
- 专业核心教材系列Professional core textbook series
- 中英双语教师实施指南Bilingual faculty implementation guide
- 更多行业与硬件可替换模块More industry and replaceable hardware modules
面向职业教育的AI课程与项目实践AI Curriculum and Project Practice for Vocational Education
面向职业院校、教师和教育合作伙伴，提供可阅读、可运行、可复核、可本地化的AI课程与实践资源。
AI curriculum and practice resources for vocational institutions, educators and education partners, designed to be readable, runnable, verifiable and localizable.