{"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":"metrics-and-calibration","title":"指标、阈值、校准与不确定性","english_title":"Metrics and calibration","chapter_number":41,"volume":2,"volume_title":"第二卷：经典AI与机器学习","month":11,"unit_title":"评估、因果、公平与鲁棒性","review_status":"draft","updated_at":"2026-08-06","estimated_hours":8,"focus":"根据错误代价选择指标，区分排序、分类、概率质量和置信表达。","objectives":["根据错误代价选择指标，区分排序、分类、概率质量和置信表达。","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":["self-supervised-representation"],"exercise_count":3,"tracks":["engineering","research"],"human_url":"/learn/metrics-and-calibration"},"labs":[{"id":"lab-metrics-and-calibration","chapter_slug":"metrics-and-calibration","title":"指标、阈值、校准与不确定性：单变量实验","kind":"decision_threshold","data_status":"教学模拟","variable":{"label":"决策阈值","unit":"","min":0,"max":1,"step":0.05,"default":0.5},"objective":"只改变“决策阈值”，观察结果、代价和风险如何一起变化，理解根据错误代价选择指标，区分排序、分类、概率质量和置信表达。","static_fallback":"静态替代：把决策阈值分别设为0、0.5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":["S022","S023","S024","S071","S072"]}],"project":{"id":"project-11","month":11,"title":"建立一个模型审计小组报告","brief":"对同一模型做指标、校准、群体切片、数据漂移、因果假设和对抗测试，记录不确定性。","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":"S022","author":"Scott Cunningham","title":"Causal Inference: The Mixtape","source_level":"开放教材","published_at":"2021","accessed_at":"2026-08-06","doi_or_url":"https://mixtape.scunning.com/"},{"id":"S023","author":"Fairlearn contributors","title":"Fairlearn User Guide","source_level":"开源项目官方文档","published_at":"持续更新","accessed_at":"2026-08-06","doi_or_url":"https://fairlearn.org/main/user_guide/index.html"},{"id":"S024","author":"NIST","title":"Adversarial Machine Learning: A Taxonomy and Terminology of Attacks and Mitigations","source_level":"标准机构一手资料","published_at":"2025","accessed_at":"2026-08-06","doi_or_url":"https://csrc.nist.gov/pubs/ai/100/2/e2025/final"},{"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"},{"id":"S072","author":"Timnit Gebru et al.","title":"Datasheets for Datasets","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/1803.09010"}]}