{"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":"gradient-optimizers","title":"梯度下降、动量与Adam","english_title":"Gradient optimizers","chapter_number":23,"volume":1,"volume_title":"第一卷：计算与数学地基","month":6,"unit_title":"微积分、优化和数值计算","review_status":"draft","updated_at":"2026-08-06","estimated_hours":8,"focus":"比较批量梯度、随机梯度、动量、自适应步长的假设和失败方式。","objectives":["比较批量梯度、随机梯度、动量、自适应步长的假设和失败方式。","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":["jacobian-hessian-autodiff"],"exercise_count":3,"tracks":["engineering","research"],"human_url":"/learn/gradient-optimizers"},"labs":[{"id":"lab-gradient-optimizers","chapter_slug":"gradient-optimizers","title":"梯度下降、动量与Adam：单变量实验","kind":"learning_rate","data_status":"教学模拟","variable":{"label":"学习率","unit":"","min":0.0001,"max":1,"step":0.0001,"default":0.03},"objective":"只改变“学习率”，观察结果、代价和风险如何一起变化，理解比较批量梯度、随机梯度、动量、自适应步长的假设和失败方式。","static_fallback":"静态替代：把学习率分别设为0.0001、0.03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":["S009","S014","S015","S025"]}],"project":{"id":"project-06","month":6,"title":"训练一条可视化的微型回归曲线","brief":"从损失函数、梯度、学习率到稳定性完整记录训练，比较至少三种设置。","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":"S009","author":"NumPy Developers","title":"NumPy Documentation","source_level":"官方文档","published_at":"持续更新","accessed_at":"2026-08-06","doi_or_url":"https://numpy.org/doc/stable/"},{"id":"S014","author":"Stephen Boyd, Lieven Vandenberghe","title":"Convex Optimization","source_level":"大学开放教材","published_at":"2004","accessed_at":"2026-08-06","doi_or_url":"https://web.stanford.edu/~boyd/cvxbook/"},{"id":"S015","author":"PyTorch","title":"Automatic Differentiation with torch.autograd","source_level":"官方教程","published_at":"持续更新","accessed_at":"2026-08-06","doi_or_url":"https://pytorch.org/tutorials/beginner/basics/autogradqs_tutorial.html"},{"id":"S025","author":"PyTorch","title":"PyTorch Tutorials","source_level":"官方教程","published_at":"持续更新","accessed_at":"2026-08-06","doi_or_url":"https://pytorch.org/tutorials/"}]}