experiment-design

Compare original and translation side by side

🇺🇸

Original

English

🇨🇳

Translation

Chinese

Experiment Design

实验设计

Design structured, progressive experiment plans for research papers.

为科研论文设计结构化、递进式的实验方案。

Input

输入

```
$0
```
— Research idea, plan, or method description

```
$0
```
— 研究想法、规划或方法描述

References

参考资料

4-stage progressive experiment prompts:

~/.claude/skills/experiment-design/references/stage-prompts.md

四阶段递进式实验提示词：

~/.claude/skills/experiment-design/references/stage-prompts.md

Scripts

脚本

Generate experiment design

生成实验设计

bash

python ~/.claude/skills/experiment-design/scripts/design_experiments.py --plan research_plan.json --output experiment_design.json
python ~/.claude/skills/experiment-design/scripts/design_experiments.py --method "contrastive learning" --task classification --format markdown

Generates baselines, ablation matrix, hyperparameter grid, metric selection. Stdlib-only.

bash

python ~/.claude/skills/experiment-design/scripts/design_experiments.py --plan research_plan.json --output experiment_design.json
python ~/.claude/skills/experiment-design/scripts/design_experiments.py --method "contrastive learning" --task classification --format markdown

生成基线模型、消融矩阵、超参数网格和评估指标选择。仅使用标准库实现。

4-Stage Progressive Framework (from AI-Scientist-v2)

四阶段递进式框架（源自AI-Scientist-v2）

Stage 1: Initial Implementation

阶段1：初始实现

Focus on getting a basic working implementation
Use a simple dataset
Aim for basic functional correctness
Completion: at least one working (non-buggy) implementation

重点是搭建可运行的基础实现
使用简单数据集
目标是保证基础功能正确性
完成标准：至少一个可正常运行（无bug）的实现

Stage 2: Baseline Tuning

阶段2：基线调优

Tune hyperparameters (learning rate, epochs, batch size)
Do NOT change model architecture
Test on at least TWO datasets
Completion: stable training curves, improvement over Stage 1

调优超参数（学习率、训练轮数、批次大小）
不得更改模型架构
至少在两个数据集上进行测试
完成标准：稳定的训练曲线，效果优于阶段1

Stage 3: Creative Research

阶段3：创新性研究

Explore novel improvements and insights
Be creative and think outside the box
Test on at least THREE datasets
Completion: demonstrated novel improvement

探索新颖的改进方向和研究见解
发挥创意，跳出固有思维
至少在三个数据集上进行测试
完成标准：展示出新颖的改进效果

Stage 4: Ablation Studies

阶段4：消融实验

Systematic component analysis
Each ablation tests a different aspect
Use same datasets as Stage 3
Completion: all planned ablations done

系统性的组件分析
每个消融实验针对不同的组件方面
使用与阶段3相同的数据集
完成标准：完成所有规划的消融实验

Output Format

输出格式

json

{
  "stages": [
    {
      "name": "initial_implementation",
      "goals": ["Basic working baseline", "Simple dataset"],
      "max_iterations": 5,
      "completion_criteria": "Working implementation with non-zero accuracy"
    }
  ],
  "baselines": ["Method A", "Method B"],
  "datasets": ["Dataset1", "Dataset2", "Dataset3"],
  "metrics": ["accuracy", "F1", "inference_time"],
  "ablation_components": ["component_A", "component_B"],
  "hyperparameter_grid": {
    "lr": [1e-4, 1e-3, 1e-2],
    "batch_size": [32, 64, 128]
  },
  "num_seeds": 3
}