面向对象与背景
Explain why a learned model is appropriate for the task.
需要解决的问题
Define the prediction or classification task and who benefits from it.
项目目标
- Choose meaningful baselines.
- Define evaluation criteria before experimentation.
设计约束
- Document dataset, compute, latency, fairness, or deployment constraints.
搭建的逻辑框架
- Establish a reproducible baseline.
- Run controlled experiments.
- Evaluate beyond one headline metric.
使用价值
- Report performance together with failure modes and limitations.
项目总结
- Describe the most valuable next experiment.