{"schema_version":"1.0","language":"zh-CN","generated_from":"NEWVAR 从零开始学AI structured Markdown","write_access":false,"full_text_access":false,"content_boundary":"This endpoint exposes metadata, objectives, prerequisites, labs and source summaries. It does not expose chapter body text or answer text.","chapter":{"slug":"margins-kernels-generalization","title":"间隔、核方法与泛化","english_title":"Margins, kernels and generalization","chapter_number":36,"volume":2,"volume_title":"第二卷：经典AI与机器学习","month":9,"unit_title":"监督学习","review_status":"draft","updated_at":"2026-08-06","estimated_hours":8,"focus":"从最大间隔理解SVM，用核技巧表达非线性，并把复杂度与数据量联系。","objectives":["从最大间隔理解SVM，用核技巧表达非线性，并把复杂度与数据量联系。","解释核心数学关系并完成最小实现","运行单变量实验并记录失败样例","完成工程线或研究线至少一项迁移任务"],"prerequisites":["trees-ensembles-boosting"],"exercise_count":3,"tracks":["engineering","research"],"human_url":"/learn/margins-kernels-generalization"},"labs":[{"id":"lab-margins-kernels-generalization","chapter_slug":"margins-kernels-generalization","title":"间隔、核方法与泛化：单变量实验","kind":"regularization","data_status":"教学模拟","variable":{"label":"正则强度","unit":"","min":0,"max":1,"step":0.01,"default":0.1},"objective":"只改变“正则强度”，观察结果、代价和风险如何一起变化，理解从最大间隔理解SVM，用核技巧表达非线性，并把复杂度与数据量联系。","static_fallback":"静态替代：把正则强度分别设为0、0.1和1，手工比较三组教学模拟输出。","keyboard":"聚焦滑块后使用方向键微调，Page Up/Page Down大步调整，Home/End到达边界。","measurement_note":"页面数值由公开公式生成，只用于教学，不代表真实模型性能或现实世界因果效果。","source_ids":["S020","S021","S010","S071"]}],"project":{"id":"project-09","month":9,"title":"建立一个可解释的学习资源分类器","brief":"使用虚构标签或公开数据，完成问题定义、基线、训练/验证/测试、错误分析和模型卡。","deliverables":["问题与边界说明","可运行最小实现","实验记录与失败分析","来源与许可清单","风险、隐私与人工闸门","复现README"]},"sources":[{"id":"S020","author":"scikit-learn developers","title":"scikit-learn User Guide","source_level":"官方文档","published_at":"持续更新","accessed_at":"2026-08-06","doi_or_url":"https://scikit-learn.org/stable/user_guide.html"},{"id":"S021","author":"Google","title":"Machine Learning Crash Course","source_level":"官方公开课程","published_at":"持续更新","accessed_at":"2026-08-06","doi_or_url":"https://developers.google.com/machine-learning/crash-course"},{"id":"S010","author":"Aston Zhang et al.","title":"Dive into Deep Learning","source_level":"开放教材","published_at":"持续更新","accessed_at":"2026-08-06","doi_or_url":"https://d2l.ai/"},{"id":"S071","author":"Margaret Mitchell et al.","title":"Model Cards for Model Reporting","source_level":"同行评审论文预印本","published_at":"2018","accessed_at":"2026-08-06","doi_or_url":"https://arxiv.org/abs/1810.03993"}]}