<%*
const clean = (value) => (value || "").replace(/\r?\n/g, " ").trim();
const yaml = (value) => clean(value).replace(/"/g, '\"');
const fileSafe = (value) => clean(value)
.replace(/[\/:*?"<>|#^[]]/g, "")
.replace(/\s+/g, " ")
.slice(0, 120)
.trim();
const topic = clean(await tp.system.prompt("Topic / field", tp.file.title));
const coreTask = clean(await tp.system.prompt("Core task / problem", ""));
const paperUrl = clean(await tp.system.prompt("Main leaderboard / survey / search URL", ""));
const pdf = clean(await tp.system.prompt("Main PDF wikilink or file name", ""));
const venue = clean(await tp.system.prompt("Main venue / source", ""));
const year = clean(await tp.system.prompt("Main year", ""));
const status = await tp.system.suggester(["draft", "reading", "completed"], ["draft", "reading", "completed"], false, "Status");
const noteTitle = ${topic || "Untitled"} Read Wide;
await tp.file.rename(fileSafe(noteTitle) || "Read Wide");
-%>
title: "<% yaml(noteTitle) %>"
created: <% tp.date.now("YYYY-MM-DD HH:mm") %>
type: paper-reading
status: <% status %>
tags:
- paper-reading
- read-wide
paper_title: ""
paper_url: "<% yaml(paperUrl) %>"
pdf: "<% yaml(pdf) %>"
venue: "<% yaml(venue) %>"
year: "<% yaml(year) %>"
topic: "<% yaml(topic) %>"
reading_stage: read-wide
<% noteTitle %>
阅读目标
- Topic: <% topic %>
- Core task / problem: <% coreTask %>
-
我这轮最想搞清楚的问题：
-
暂时不深入的问题：
领域地图
Task / Problem
- 这个领域想解决什么问题：
- 输入是什么：
- 输出是什么：
- 真实应用场景：
Dataset / Benchmark
常见方法分类
SOTA / Recent Papers
Paper 1
- Paper:
- Link:
- Model / method:
- Problem:
- Method:
- Result:
-
To Understand:
Paper 2
- Paper:
- Link:
- Model / method:
- Problem:
- Method:
- Result:
-
To Understand:
Paper 3
- Paper:
- Link:
- Model / method:
- Problem:
- Method:
- Result:
-
To Understand:
Dataset Papers
Dataset Paper 1
- Paper:
- Dataset:
- Data source / scale:
- Task definition:
- Metrics:
- Why this dataset matters:
-
To Understand:
Survey Papers
Survey Paper 1
- Paper:
- Link:
- Figure 1 / taxonomy:
- Contributions:
- Conclusions and Future Directions:
-
Open problems:
-
反复出现的经典 paper：
-
反复出现的 technical terms：
-
反复出现的 dataset / metric：
-
新方法主要在改什么：
Between Wide and Deep
One-paragraph Summary
用 4-6 句话总结这个 topic 的当前 mental model：
My Questions
Next Papers to Read Deep
To Understand List