AGENT SKILL

research-idea-explainer

Translate a rigorously investigated research idea into intuitive motivation, a concrete scenario, a two-case example, and a causal story a graduate student can retell. Use as part of Research Idea Discovery final delivery or for requests like 我没懂这个idea.

License 许可
MIT
最近更新
2026-10-13
Stars
★ 0
Forks
0

01 前言 Front Matter

name
research-idea-explainer
description
Translate a rigorously investigated research idea into intuitive motivation, a concrete scenario, a two-case example, and a causal story a graduate student can retell. Use as part of Research Idea Discovery final delivery or for requests like 我没懂这个idea.

03 SKILL.md 正文

Idea Explanation Layer — 导师先让学生听懂

Read ../research-idea-discovery/references/plain-language-idea-protocol.md and ../research-idea-discovery/references/scientific-story-and-contributions.md.

Output, in this order:

  1. One-sentence daily-language question (≤50 Chinese characters when practical).
  2. Concrete task sample: exactly what inputs, classes/labels, and output the old model uses. If no grounded case is available, mark it a hypothetical teaching example.
  3. Two contrast situations holding one superficial property fixed (e.g. equally confident) but varying hypothesized task evidence. State what is observed versus hypothesized; do not invent measured results.
  4. What a competent ordinary baseline does, why it may fail in the selected case, and the most dangerous simple alternative that might already solve it.
  5. Core difference in one sentence, followed by technical terms/possible method—not before.
  6. A 30-second spoken scientific story and a falsifying E0.

Forbidden: empty slogans ("confidence ≠ evidence") without operational definitions; naming a graph without an estimable edge definition; claiming baselines actually fail without evidence. If the user says 没懂, make a smaller toy example and walk through concrete inputs and prediction changes instead of repeating vocabulary.

04 同仓库更多 Skills