一个环形图形摘要提示词：为闭环系统画 GA
来源：GitHub 仓库 XYD-GIS/paper-toolkits 文件约 90 行，约 4 KB
一个环形图形摘要提示词。
它解决的是：闭环系统（强化学习、在线学习、闭环控制、生物反馈循环）要画图形摘要。
它的结构：
- 布局：大圆（占画面 90%）分成 N 个等分扇形
- 每片包含：外缘加粗 Arial 标签（2 词）、中间扁平矢量图标、内缘斜体灰色短 caption（5 词）
- 中心：小圆形徽章，加粗 Arial 论文标题（最多 10 词）
- 扇形之间：沿外缘的曲线箭头，顺时针，表示循环方向
- 颜色分配（从 12 点顺时针）：软蓝 #D6E4F0、软绿 #D8E8D0、软橙 #F5E0CB、软紫 #E0D4EC
- 底部：小号灰色斜体一句话结论
- 风格：扁平矢量、Arial、粉彩、白底、生物/系统论文审美、1:1
- 负向约束：不要照片、不要 3D、不要卡通、不要投影、不要 emoji、不要长句、不要难读文字
填空示例（强化学习闭环）：
- 阶段数量 = four
- 论文标题 = Closed-Loop Adaptive Control with Reinforcement Learning
- 结论一句话 = Each iteration refines the policy using feedback from environment rewards.
- 每片具体内容：Observe / Decide / Act / Learn
调优提示：
- 阶段数 = 3 → 三角形布局，每片占 120 度
- 阶段数 ≥ 6 → 扇片会很窄，文字易溢出，考虑用 narrative.md 横向布局替代
- 想突出某一片 → 在 prompt 中加说明
适合谁：
- 做强化学习、在线学习、闭环控制、生物反馈循环方向的科研小组成员
- 写论文要画图形摘要的人
- 想要闭环系统视觉表达的人
用法： 把这段提示词复制到 AI 工具里，把占位符填上就行。
提示词
A circular graphical abstract for an academic paper, arranged as a clockwise loop of {阶段数量，如 four} stages around the center.
LAYOUT: A large circle (90% of frame) divided into {阶段数量} equal pie slices. Each slice contains:
- A 2-word English label in bold Arial at the outer edge
- A simple flat-vector icon in the middle of the slice
- A short 5-word caption in italic gray text at the inner edge
CENTER OF CIRCLE: A small circular badge with the paper title in bold Arial (max 10 words): "{论文标题}"
ARROWS BETWEEN SLICES: Curved arrows along the outer edge of the circle, going clockwise, indicating the loop direction.
COLOR ASSIGNMENT (clockwise from 12 o'clock):
- Slice 1: soft blue #D6E4F0
- Slice 2: soft green #D8E8D0
- Slice 3: soft orange #F5E0CB
- Slice 4: soft purple #E0D4EC
(repeat or extend palette if num_stages > 4)
At the bottom of the figure, in small gray italic text, add the one-sentence takeaway: "{结论一句话}"
Style: flat vector, Arial sans-serif, pastel palette, white background, biology / systems-paper aesthetic. Aspect ratio 1:1.
Negative constraints: NO photorealistic, NO 3D, NO cartoon, NO drop shadows, NO emoji, NO long sentences, NO unreadable text.
# 填空示例
{阶段数量} = four
{论文标题} = Closed-Loop Adaptive Control with Reinforcement Learning
{结论一句话} = Each iteration refines the policy using feedback from environment rewards.
# 每片的具体内容（在 prompt 末尾追加）
SLICE DETAILS:
- Slice 1 (top, 12 o'clock): label "Observe", icon = an eye, caption "Sense environment state."
- Slice 2 (right, 3 o'clock): label "Decide", icon = a brain or compass, caption "Compute next action."
- Slice 3 (bottom, 6 o'clock): label "Act", icon = a hand pressing a button, caption "Execute in environment."
- Slice 4 (left, 9 o'clock): label "Learn", icon = a gear / arrow loop, caption "Update policy from reward."评论 0
更多
登录后可点赞、收藏、评论和举报。
还没有评论，先发起一个具体问题。