ml-paper-writing
Compare original and translation side by side
🇺🇸
Original
English🇨🇳
Translation
ChineseML Paper Writing for Top AI Conferences
面向AI顶会的ML论文撰写指南
Expert-level guidance for writing publication-ready papers targeting NeurIPS, ICML, ICLR, ACL, AAAI, and COLM. This skill combines writing philosophy from top researchers (Nanda, Farquhar, Karpathy, Lipton, Steinhardt) with practical tools: LaTeX templates, citation verification APIs, and conference checklists.
本指南提供专家级指导,帮助撰写可投稿至NeurIPS、ICML、ICLR、ACL、AAAI和COLM的论文。它融合了顶尖研究者(Nanda、Farquhar、Karpathy、Lipton、Steinhardt)的写作理念,搭配实用工具:LaTeX模板、引用校验API以及会议提交清单。
Core Philosophy: Collaborative Writing
核心理念:协作式写作
Paper writing is collaborative, but Claude should be proactive in delivering drafts.
The typical workflow starts with a research repository containing code, results, and experimental artifacts. Claude's role is to:
- Understand the project by exploring the repo, results, and existing documentation
- Deliver a complete first draft when confident about the contribution
- Search literature using web search and APIs to find relevant citations
- Refine through feedback cycles when the scientist provides input
- Ask for clarification only when genuinely uncertain about key decisions
Key Principle: Be proactive. If the repo and results are clear, deliver a full draft. Don't block waiting for feedback on every section—scientists are busy. Produce something concrete they can react to, then iterate based on their response.
论文写作是协作过程,但Claude应主动产出草稿。
典型工作流从包含代码、结果和实验工件的研究仓库开始。Claude的职责是:
- 理解项目:探索仓库、结果和现有文档
- 交付完整初稿:在明确核心贡献后产出完整初稿
- 检索文献:通过网页搜索和API查找相关引用
- 基于反馈迭代:根据科研人员的输入优化内容
- 仅在必要时请求澄清:仅对关键决策存在真实不确定性时才询问
关键原则:主动出击。如果仓库和结果清晰明确,直接交付完整草稿。不要因等待每部分的反馈而停滞——科研人员时间宝贵。先产出具体内容供他们反馈,再根据回应迭代。
⚠️ CRITICAL: Never Hallucinate Citations
⚠️ 重中之重:绝不能虚构引用
This is the most important rule in academic writing with AI assistance.
这是AI辅助学术写作中最重要的规则。
The Problem
问题所在
AI-generated citations have a ~40% error rate. Hallucinated references—papers that don't exist, wrong authors, incorrect years, fabricated DOIs—are a serious form of academic misconduct that can result in desk rejection or retraction.
AI生成的引用错误率约为40%。虚构的参考文献——不存在的论文、错误的作者、年份不符、编造的DOI——属于严重学术不端行为,可能导致直接拒稿或撤稿。
The Rule
规则
NEVER generate BibTeX entries from memory. ALWAYS fetch programmatically.
| Action | ✅ Correct | ❌ Wrong |
|---|---|---|
| Adding a citation | Search API → verify → fetch BibTeX | Write BibTeX from memory |
| Uncertain about a paper | Mark as | Guess the reference |
| Can't find exact paper | Note: "placeholder - verify" | Invent similar-sounding paper |
绝不能凭记忆生成BibTeX条目。必须通过程序化方式获取。
| 操作 | ✅ 正确做法 | ❌ 错误做法 |
|---|---|---|
| 添加引用 | 搜索API → 验证 → 获取BibTeX | 凭记忆编写BibTeX |
| 对某篇论文存疑 | 标记为 | 猜测参考文献 |
| 无法找到精确匹配的论文 | 标注:"placeholder - verify" | 编造名称相似的论文 |
When You Can't Verify a Citation
无法验证引用时的处理
If you cannot programmatically verify a citation, you MUST:
latex
% EXPLICIT PLACEHOLDER - requires human verification
\cite{PLACEHOLDER_author2024_verify_this} % TODO: Verify this citation existsAlways tell the scientist: "I've marked [X] citations as placeholders that need verification. I could not confirm these papers exist."
如果无法通过程序化方式验证引用,必须:
latex
% 明确占位符 - 需要人工验证
\cite{PLACEHOLDER_author2024_verify_this} % TODO: 请验证此引用是否存在务必告知科研人员:"我已将[X]处引用标记为需要验证的占位符,无法确认这些论文是否存在。"
Recommended: Install Exa MCP for Paper Search
推荐:安装Exa MCP用于论文检索
For the best paper search experience, install Exa MCP which provides real-time academic search:
Claude Code:
bash
claude mcp add exa -- npx -y mcp-remote "https://mcp.exa.ai/mcp"Cursor / VS Code (add to MCP settings):
json
{
"mcpServers": {
"exa": {
"type": "http",
"url": "https://mcp.exa.ai/mcp"
}
}
}Exa MCP enables searches like:
- "Find papers on RLHF for language models published after 2023"
- "Search for transformer architecture papers by Vaswani"
- "Get recent work on sparse autoencoders for interpretability"
Then verify results with Semantic Scholar API and fetch BibTeX via DOI.
为获得最佳论文检索体验,请安装Exa MCP,它提供实时学术搜索功能:
Claude Code命令:
bash
claude mcp add exa -- npx -y mcp-remote "https://mcp.exa.ai/mcp"Cursor / VS Code(添加至MCP设置):
json
{
"mcpServers": {
"exa": {
"type": "http",
"url": "https://mcp.exa.ai/mcp"
}
}
}Exa MCP支持以下类型的搜索:
- "查找2023年后发表的关于语言模型RLHF的论文"
- "检索Vaswani关于Transformer架构的论文"
- "获取稀疏自编码器可解释性的最新研究"
随后可通过Semantic Scholar API验证并获取BibTeX。
Workflow 0: Starting from a Research Repository
工作流0:从研究仓库起步
When beginning paper writing, start by understanding the project:
Project Understanding:
- [ ] Step 1: Explore the repository structure
- [ ] Step 2: Read README, existing docs, and key results
- [ ] Step 3: Identify the main contribution with the scientist
- [ ] Step 4: Find papers already cited in the codebase
- [ ] Step 5: Search for additional relevant literature
- [ ] Step 6: Outline the paper structure together
- [ ] Step 7: Draft sections iteratively with feedbackStep 1: Explore the Repository
bash
undefined开始撰写论文时,先从理解项目着手:
项目理解步骤:
- [ ] 步骤1:探索仓库结构
- [ ] 步骤2:阅读README、现有文档和关键结果
- [ ] 步骤3:与科研人员确认核心贡献
- [ ] 步骤4:查找代码库中已有的引用论文
- [ ] 步骤5:搜索额外的相关文献
- [ ] 步骤6:共同规划论文结构
- [ ] 步骤7:基于反馈迭代撰写各章节步骤1:探索仓库
bash
undefinedUnderstand project structure
了解项目结构
ls -la
find . -name ".py" | head -20
find . -name ".md" -o -name "*.txt" | xargs grep -l -i "result|conclusion|finding"
Look for:
- `README.md` - Project overview and claims
- `results/`, `outputs/`, `experiments/` - Key findings
- `configs/` - Experimental settings
- Existing `.bib` files or citation references
- Any draft documents or notes
**Step 2: Identify Existing Citations**
Check for papers already referenced in the codebase:
```bashls -la
find . -name ".py" | head -20
find . -name ".md" -o -name "*.txt" | xargs grep -l -i "result|conclusion|finding"
重点关注:
- `README.md` - 项目概述和核心论点
- `results/`、`outputs/`、`experiments/` - 关键发现
- `configs/` - 实验设置
- 现有的`.bib`文件或引用参考
- 任何草稿文档或笔记
**步骤2:识别现有引用**
检查代码库中已引用的论文:
```bashFind existing citations
查找现有引用
grep -r "arxiv|doi|cite" --include=".md" --include=".bib" --include=".py"
find . -name ".bib"
These are high-signal starting points for Related Work—the scientist has already deemed them relevant.
**Step 3: Clarify the Contribution**
Before writing, explicitly confirm with the scientist:
> "Based on my understanding of the repo, the main contribution appears to be [X].
> The key results show [Y]. Is this the framing you want for the paper,
> or should we emphasize different aspects?"
**Never assume the narrative—always verify with the human.**
**Step 4: Search for Additional Literature**
Use web search to find relevant papers:
Search queries to try:
- "[main technique] + [application domain]"
- "[baseline method] comparison"
- "[problem name] state-of-the-art"
- Author names from existing citations
Then verify and retrieve BibTeX using the citation workflow below.
**Step 5: Deliver a First Draft**
**Be proactive—deliver a complete draft rather than asking permission for each section.**
If the repo provides clear results and the contribution is apparent:
1. Write the full first draft end-to-end
2. Present the complete draft for feedback
3. Iterate based on scientist's response
If genuinely uncertain about framing or major claims:
1. Draft what you can confidently
2. Flag specific uncertainties: "I framed X as the main contribution—let me know if you'd prefer to emphasize Y instead"
3. Continue with the draft rather than blocking
**Questions to include with the draft** (not before):
- "I emphasized X as the main contribution—adjust if needed"
- "I highlighted results A, B, C—let me know if others are more important"
- "Related work section includes [papers]—add any I missed"
---grep -r "arxiv|doi|cite" --include=".md" --include=".bib" --include=".py"
find . -name ".bib"
这些是相关工作部分的高信号起点——科研人员已认定它们与项目相关。
**步骤3:明确核心贡献**
撰写前,需与科研人员明确确认:
> "根据我对仓库的理解,核心贡献似乎是[X]。
> 关键结果显示[Y]。这是否是你希望的论文框架,
> 还是我们需要强调不同的方面?"
**绝不要主观臆断叙事方向——务必与人工确认。**
**步骤4:搜索额外文献**
使用网页搜索查找相关论文:
可尝试的搜索查询:
- "[核心技术] + [应用领域]"
- "[基线方法] 对比"
- "[问题名称] 最新进展"
- 现有引用中的作者姓名
随后按照下方的引用工作流进行验证和获取BibTeX。
**步骤5:交付初稿**
**主动出击——交付完整草稿,而非逐节请求许可。**
如果仓库提供了清晰的结果且核心贡献明确:
1. 从头到尾撰写完整的初稿
2. 提交完整草稿供反馈
3. 根据科研人员的回应迭代
如果对框架或主要论点确实存在疑问:
1. 先撰写你有把握的部分
2. 标记具体的不确定性:"我将X作为核心贡献进行阐述——如果您希望强调Y,请告知我"
3. 继续撰写草稿,不要停滞
**随草稿一并提出问题**(而非提前询问):
- "我将X作为核心贡献重点阐述——如有需要可调整"
- "我突出了结果A、B、C——若有更重要的结果请告知"
- "相关工作部分引用了[论文]——如有遗漏请补充"
---When to Use This Skill
适用场景
Use this skill when:
- Starting from a research repo to write a paper
- Drafting or revising specific sections
- Conducting literature reviews and finding related work
- Discovering recent papers in your research area
- Finding and verifying citations for related work
- Formatting for conference submission
- Resubmitting to a different venue (format conversion)
- Iterating on drafts with scientist feedback
Always remember: First drafts are starting points for discussion, not final outputs.
在以下场景中使用本技能:
- 从研究仓库起步撰写论文
- 起草或修订特定章节
- 开展文献综述并查找相关研究
- 发现研究领域的最新论文
- 查找并校验相关工作的引用
- 格式调整以符合会议提交要求
- 转投至其他会议(格式转换)
- 基于科研人员反馈迭代草稿
请始终牢记:初稿是讨论的起点,而非最终成果。
Workflow: Literature Research & Paper Discovery
工作流:文献调研与论文检索
When conducting literature reviews, finding related work, or discovering recent papers, use this workflow to systematically search, evaluate, and select ML papers.
开展文献综述、查找相关研究或发现最新论文时,使用以下工作流系统地搜索、评估和筛选ML论文。
Workflow 5: Finding and Evaluating Papers
工作流5:论文查找与评估
Literature Research Process:
- [ ] Step 1: Define search scope and keywords
- [ ] Step 2: Search arXiv and academic databases
- [ ] Step 3: Screen papers by title/abstract
- [ ] Step 4: Evaluate paper quality (5 dimensions)
- [ ] Step 5: Select top papers and extract citations
- [ ] Step 6: Verify citations programmaticallyStep 1: Define Search Scope
Identify specific research areas, methods, or applications:
- Technique-focused: ,
transformer architecture,graph neural networksself-supervised learning - Application-focused: ,
medical image analysis,reinforcement learning for roboticslanguage model alignment - Problem-focused: ,
out-of-distribution generalization,continual learningfairness in ML
Step 2: Search arXiv
Use arXiv search with targeted keywords:
URL Pattern:
https://arxiv.org/search/?searchtype=all&query=KEYWORDS&abstracts=show&order=-announced_date_first
Example Searches:
- https://arxiv.org/search/?searchtype=all&query=graph+neural+networks&abstracts=show&order=-announced_date_first
- https://arxiv.org/search/?cat:cs.LG+AND+all:transformer&abstracts=show&order=-announced_date_firstTips:
- Combine keywords with for AND
+ - Filter by categories: ,
cs.LG,cs.AI,cs.CVcs.CL - Sort by for recent papers
announced_date_first - Use Chrome MCP tools when available for automation
Step 3: Screen Papers
Quick screening by title and abstract:
- Relevance to research topic
- Novelty of contribution
- Venue/reputation of authors
- Code availability (check for GitHub links)
Step 4: Evaluate Quality
Use the 5-dimension quality criteria:
| Dimension | Weight | Evaluation Focus |
|---|---|---|
| Innovation | 30% | Novelty and originality |
| Method Completeness | 25% | Clarity and reproducibility |
| Experimental Thoroughness | 25% | Validation depth |
| Writing Quality | 10% | Presentation clarity |
| Relevance & Impact | 10% | Domain importance |
Scoring: Rate each dimension 1-5, calculate weighted total
Step 5: Select and Extract
- Rank papers by total score
- Select top papers for detailed review
- Extract metadata: title, authors, arXiv ID, abstract
- Note code repository links
Step 6: Verify Citations
For selected papers, verify citations using Semantic Scholar API:
- Fetch BibTeX programmatically via DOI
- Mark unverified citations as
[CITATION NEEDED] - Store in bibliography with verification status
文献调研流程:
- [ ] 步骤1:定义搜索范围和关键词
- [ ] 步骤2:检索arXiv和学术数据库
- [ ] 步骤3:通过标题/摘要筛选论文
- [ ] 步骤4:从5个维度评估论文质量
- [ ] 步骤5:筛选优质论文并提取引用
- [ ] 步骤6:程序化验证引用步骤1:定义搜索范围
明确具体的研究领域、方法或应用方向:
- 技术导向:、
transformer architecture、graph neural networksself-supervised learning - 应用导向:、
medical image analysis、reinforcement learning for roboticslanguage model alignment - 问题导向:、
out-of-distribution generalization、continual learningfairness in ML
步骤2:检索arXiv
使用arXiv搜索并搭配目标关键词:
URL模式:
https://arxiv.org/search/?searchtype=all&query=KEYWORDS&abstracts=show&order=-announced_date_first
示例搜索:
- https://arxiv.org/search/?searchtype=all&query=graph+neural+networks&abstracts=show&order=-announced_date_first
- https://arxiv.org/search/?cat:cs.LG+AND+all:transformer&abstracts=show&order=-announced_date_first技巧:
- 使用组合关键词表示逻辑与
+ - 通过分类筛选:、
cs.LG、cs.AI、cs.CVcs.CL - 按排序获取最新论文
announced_date_first - 若可用,使用Chrome MCP工具实现自动化
步骤3:筛选论文
通过标题和摘要快速筛选:
- 与研究主题的相关性
- 贡献的创新性
- 作者的会议/声誉
- 代码可用性(检查是否有GitHub链接)
步骤4:质量评估
使用5维度质量评估标准:
| 维度 | 权重 | 评估重点 |
|---|---|---|
| 创新性 | 30% | 新颖性和原创性 |
| 方法完整性 | 25% | 清晰度和可复现性 |
| 实验严谨性 | 25% | 验证深度 |
| 写作质量 | 10% | 表述清晰度 |
| 相关性与影响力 | 10% | 领域重要性 |
评分方式:每个维度按1-5分评分,计算加权总分
步骤5:筛选与提取
- 按总分对论文排序
- 选择排名靠前的论文进行详细综述
- 提取元数据:标题、作者、arXiv ID、摘要
- 记录代码仓库链接
步骤6:验证引用
对于筛选出的论文,使用Semantic Scholar API验证引用:
- 通过DOI程序化获取BibTeX
- 将未验证的引用标记为
[CITATION NEEDED] - 按验证状态存储在参考文献中
When to Use Literature Research
文献调研的适用场景
Use this workflow when:
- Starting a new project: Find related work and baselines
- Writing Related Work section: Discover recent papers in your area
- Staying updated: Track recent publications in your field
- Finding baselines: Identify state-of-the-art methods for comparison
- Literature review: Comprehensive survey of research area
在以下场景中使用本工作流:
- 启动新项目:查找相关研究和基线方法
- 撰写相关工作章节:发现领域内的最新论文
- 保持跟进:追踪领域内的最新发表成果
- 查找基线方法:识别用于对比的当前最优方法
- 文献综述:对研究领域进行全面调研
Quality Thresholds
质量阈值
- Excellent: 4.0+ (include definitely)
- Good: 3.5-3.9 (include if relevant)
- Fair: 3.0-3.4 (include if highly relevant)
- Poor: <3.0 (exclude unless essential)
- 优秀:4.0分及以上(务必纳入)
- 良好:3.5-3.9分(相关则纳入)
- 一般:3.0-3.4分(高度相关则纳入)
- 较差:低于3.0分(除非必要否则排除)
Reference Files
参考文件
For detailed literature research guidance:
- - arXiv search strategies and URL patterns
references/literature-research/arxiv-search-guide.md - - Detailed 5-dimension evaluation rubrics
references/literature-research/paper-quality-criteria.md
如需详细的文献调研指导:
- - arXiv搜索策略和URL模式
references/literature-research/arxiv-search-guide.md - - 详细的5维度评估细则
references/literature-research/paper-quality-criteria.md
Knowledge Base: Writing Patterns from ML Papers
知识库:ML论文写作模式
This skill maintains a curated knowledge base of writing patterns, techniques, and requirements extracted from successful ML conference papers. The knowledge base grows as you analyze more papers.
本技能维护一个精心整理的知识库,包含从成功的ML会议论文中提取的写作模式、技巧和要求。随着分析的论文增多,知识库会不断扩充。
Knowledge Organization
知识组织
The knowledge base is organized into 4 categories at :
references/knowledge/| Category | File | Contents |
|---|---|---|
| Structure | | Paper organization, IMRaD patterns, transitions, section flow |
| Writing Techniques | | Sentence patterns, transition phrases, clarity techniques |
| Submission Guides | | Venue requirements (NeurIPS, ICML, ICLR, ACL, AAAI, COLM) |
| Review Response | | Rebuttal strategies, addressing reviewer comments |
知识库分为4个类别,存储于目录下:
references/knowledge/| 类别 | 文件 | 内容 |
|---|---|---|
| 结构 | | 论文组织结构、IMRaD模式、过渡方式、章节逻辑 |
| 写作技巧 | | 句式模式、过渡短语、清晰度提升技巧 |
| 提交指南 | | 各会议要求(NeurIPS、ICML、ICLR、ACL、AAAI、COLM) |
| 审稿回复 | | 反驳策略、回应审稿人意见的方法 |
How the Knowledge Base is Maintained
知识库维护方式
The paper-miner agent automatically extracts and categorizes writing knowledge from papers you provide:
You: "Learn writing techniques from this NeurIPS paper: path/to/paper.pdf"
↓
paper-miner analyzes the paper
↓
Extracts patterns → Categorizes into 4 types → Updates knowledge files
↓
Knowledge grows with each paper analyzedWhat gets extracted:
- Structure patterns: How successful papers organize sections, transition between topics
- Writing techniques: Sentence templates, transition phrases, clarity methods
- Venue requirements: Page limits, required sections, formatting rules
- Rebuttal strategies: How to respond to specific reviewer concerns
paper-miner代理会自动从你提供的论文中提取并分类写作知识:
你:"从这篇NeurIPS论文中学习写作技巧:path/to/paper.pdf"
↓
paper-miner分析论文
↓
提取模式 → 分类为4种类型 → 更新知识文件
↓
知识库随每篇分析的论文不断扩充提取内容包括:
- 结构模式:成功论文如何组织章节、实现主题过渡
- 写作技巧:句式模板、过渡短语、清晰度方法
- 会议要求:页数限制、必填章节、格式规则
- 反驳策略:如何回应特定的审稿人关注点
When to Use the Knowledge Base
知识库的使用场景
For writing patterns:
- Stuck on how to phrase a transition? Check
writing-techniques.md - Need structure inspiration? Browse
structure.md - Writing rebuttal? Consult
review-response.md
For venue requirements:
- Submitting to NeurIPS? See for checklist
submission-guides.md - Converting between venues? Compare page limits and requirements
- Unsure about required sections? Each venue has specific requirements
查找写作模式时:
- 不知如何撰写过渡句?查看
writing-techniques.md - 需要结构灵感?浏览
structure.md - 撰写反驳信?参考
review-response.md
了解会议要求时:
- 投稿至NeurIPS?查看中的清单
submission-guides.md - 转投其他会议?对比页数限制和要求
- 不确定必填章节?每个会议都有具体要求
Contributing to the Knowledge Base
贡献知识库
Every paper you analyze makes the knowledge base richer for future use:
bash
undefined你分析的每一篇论文都会让知识库对未来的使用更有价值:
bash
undefinedTrigger paper-miner from any context
在任何上下文触发paper-miner
"Extract writing patterns from this paper: path/to/paper.pdf"
"Analyze structure of https://arxiv.org/abs/2301.xxxxx"
"What writing techniques does this ICLR paper use?"
The paper-miner agent:
1. Extracts paper content (PDF, DOCX, or arXiv link)
2. Analyzes IMRaD structure and writing patterns
3. Identifies venue-specific requirements
4. Updates appropriate knowledge files with new patterns
5. Reports what was added with source attribution"从这篇论文中提取写作模式:path/to/paper.pdf"
"分析https://arxiv.org/abs/2301.xxxxx的结构"
"这篇ICLR论文使用了哪些写作技巧?"
paper-miner代理会:
1. 提取论文内容(PDF、DOCX或arXiv链接)
2. 分析IMRaD结构和写作模式
3. 识别会议特定要求
4. 将新模式更新至对应知识文件
5. 报告新增内容及来源Knowledge Base Principles
知识库原则
Actionable patterns only: Each entry provides reusable techniques with examples.
Source attribution: Every pattern cites the paper it came from for traceability.
No duplicates: Checks existing content before adding new patterns.
Quality over quantity: Focus on techniques that work, not comprehensive lists.
See for complete knowledge base documentation.
references/knowledge/README.md仅收录可操作的模式:每条条目都提供可复用的技巧及示例。
来源归因:每个模式都标注其来源论文,确保可追溯。
无重复内容:添加新模式前先检查现有内容。
质量优先于数量:聚焦实用技巧,而非追求全面性。
完整的知识库文档请见。
references/knowledge/README.mdBalancing Proactivity and Collaboration
平衡主动性与协作性
Default: Be proactive. Deliver drafts, then iterate.
| Confidence Level | Action |
|---|---|
| High (clear repo, obvious contribution) | Write full draft, deliver, iterate on feedback |
| Medium (some ambiguity) | Write draft with flagged uncertainties, continue |
| Low (major unknowns) | Ask 1-2 targeted questions, then draft |
Draft first, ask with the draft (not before):
| Section | Draft Autonomously | Flag With Draft |
|---|---|---|
| Abstract | Yes | "Framed contribution as X—adjust if needed" |
| Introduction | Yes | "Emphasized problem Y—correct if wrong" |
| Methods | Yes | "Included details A, B, C—add missing pieces" |
| Experiments | Yes | "Highlighted results 1, 2, 3—reorder if needed" |
| Related Work | Yes | "Cited papers X, Y, Z—add any I missed" |
Only block for input when:
- Target venue is unclear (affects page limits, framing)
- Multiple contradictory framings seem equally valid
- Results seem incomplete or inconsistent
- Explicit request to review before continuing
Don't block for:
- Word choice decisions
- Section ordering
- Which specific results to show (make a choice, flag it)
- Citation completeness (draft with what you find, note gaps)
默认原则:主动出击。先交付草稿,再迭代。
| 信心程度 | 行动 |
|---|---|
| 高(仓库清晰,贡献明确) | 撰写完整草稿,交付后根据反馈迭代 |
| 中(存在部分模糊点) | 撰写草稿并标记不确定性,继续推进 |
| 低(存在重大未知) | 提出1-2个针对性问题,然后起草 |
先写草稿,随草稿一并提问(而非提前询问):
| 章节 | 自主撰写 | 随草稿标记 |
|---|---|---|
| 摘要 | 是 | "将贡献框架定为X——如有需要可调整" |
| 引言 | 是 | "强调了问题Y——若有误请纠正" |
| 方法 | 是 | "包含了细节A、B、C——如有遗漏请补充" |
| 实验 | 是 | "突出了结果1、2、3——如有需要可重新排序" |
| 相关工作 | 是 | "引用了论文X、Y、Z——如有遗漏请补充" |
仅在以下情况时暂停等待输入:
- 目标会议不明确(影响页数限制和框架)
- 存在多个看似同样合理的矛盾框架
- 结果似乎不完整或不一致
- 明确要求在继续前先进行审核
无需暂停的情况:
- 措辞选择
- 章节排序
- 展示哪些具体结果(自行选择并标记)
- 引用完整性(先使用已找到的内容起草,标注缺口)
The Narrative Principle
叙事原则
The single most critical insight: Your paper is not a collection of experiments—it's a story with one clear contribution supported by evidence.
Every successful ML paper centers on what Neel Nanda calls "the narrative": a short, rigorous, evidence-based technical story with a takeaway readers care about.
Three Pillars (must be crystal clear by end of introduction):
| Pillar | Description | Example |
|---|---|---|
| The What | 1-3 specific novel claims within cohesive theme | "We prove that X achieves Y under condition Z" |
| The Why | Rigorous empirical evidence supporting claims | Strong baselines, experiments distinguishing hypotheses |
| The So What | Why readers should care | Connection to recognized community problems |
If you cannot state your contribution in one sentence, you don't yet have a paper.
最关键的洞察:你的论文不是实验的堆砌——它是一个故事,围绕一个清晰的核心贡献,由证据支撑。
每篇成功的ML论文都围绕Neel Nanda所说的"叙事"展开:一个简短、严谨、基于证据的技术故事,包含读者关心的核心观点。
三大支柱(引言结尾必须清晰呈现):
| 支柱 | 描述 | 示例 |
|---|---|---|
| 是什么 | 1-3个具体、新颖且主题一致的论点 | "我们证明X在条件Z下可实现Y" |
| 为什么 | 支撑论点的严谨实证证据 | 强大的基线、区分假设的实验 |
| 重要性 | 读者为何需要关注 | 与社区公认问题的关联 |
如果你无法用一句话概括核心贡献,说明你还没有形成完整的论文框架。
Paper Structure Workflow
论文结构工作流
Workflow 1: Writing a Complete Paper (Iterative)
工作流1:完整论文撰写(迭代式)
Copy this checklist and track progress. Each step involves drafting → feedback → revision:
Paper Writing Progress:
- [ ] Step 1: Define the one-sentence contribution (with scientist)
- [ ] Step 2: Draft Figure 1 → get feedback → revise
- [ ] Step 3: Draft abstract → get feedback → revise
- [ ] Step 4: Draft introduction → get feedback → revise
- [ ] Step 5: Draft methods → get feedback → revise
- [ ] Step 6: Draft experiments → get feedback → revise
- [ ] Step 7: Draft related work → get feedback → revise
- [ ] Step 8: Draft limitations → get feedback → revise
- [ ] Step 9: Complete paper checklist (required)
- [ ] Step 10: Final review cycle and submissionStep 1: Define the One-Sentence Contribution
This step requires explicit confirmation from the scientist.
Before writing anything, articulate and verify:
- What is the single thing your paper contributes?
- What was not obvious or present before your work?
"I propose framing the contribution as: '[one sentence]'. Does this capture what you see as the main takeaway? Should we adjust the emphasis?"
Step 2: Draft Figure 1
Figure 1 deserves special attention—many readers skip directly to it.
- Convey core idea, approach, or most compelling result
- Use vector graphics (PDF/EPS for plots)
- Write captions that stand alone without main text
- Ensure readability in black-and-white (8% of men have color vision deficiency)
Step 3: Write Abstract (5-Sentence Formula)
From Sebastian Farquhar (DeepMind):
1. What you achieved: "We introduce...", "We prove...", "We demonstrate..."
2. Why this is hard and important
3. How you do it (with specialist keywords for discoverability)
4. What evidence you have
5. Your most remarkable number/resultDelete generic openings like "Large language models have achieved remarkable success..."
Step 4: Write Introduction (1-1.5 pages max)
Must include:
- 2-4 bullet contribution list (max 1-2 lines each in two-column format)
- Clear problem statement
- Brief approach overview
- Methods should start by page 2-3 maximum
Step 5: Methods Section
Enable reimplementation:
- Conceptual outline or pseudocode
- All hyperparameters listed
- Architectural details sufficient for reproduction
- Present final design decisions; ablations go in experiments
Step 6: Experiments Section
For each experiment, explicitly state:
- What claim it supports
- How it connects to main contribution
- Experimental setting (details in appendix)
- What to observe: "the blue line shows X, which demonstrates Y"
Requirements:
- Error bars with methodology (standard deviation vs standard error)
- Hyperparameter search ranges
- Compute infrastructure (GPU type, total hours)
- Seed-setting methods
Step 7: Related Work
Organize methodologically, not paper-by-paper:
Good: "One line of work uses Floogledoodle's assumption [refs] whereas we use Doobersnoddle's assumption because..."
Bad: "Snap et al. introduced X while Crackle et al. introduced Y."
Cite generously—reviewers likely authored relevant papers.
Step 8: Limitations Section (REQUIRED)
All major conferences require this. Counter-intuitively, honesty helps:
- Reviewers are instructed not to penalize honest limitation acknowledgment
- Pre-empt criticisms by identifying weaknesses first
- Explain why limitations don't undermine core claims
Step 9: Paper Checklist
NeurIPS, ICML, and ICLR all require paper checklists. See references/checklists.md.
复制此清单并跟踪进度。每个步骤都包含起草 → 反馈 → 修订:
论文撰写进度:
- [ ] 步骤1:与科研人员确定核心贡献(一句话概括)
- [ ] 步骤2:起草图1 → 获取反馈 → 修订
- [ ] 步骤3:起草摘要 → 获取反馈 → 修订
- [ ] 步骤4:起草引言 → 获取反馈 → 修订
- [ ] 步骤5:起草方法 → 获取反馈 → 修订
- [ ] 步骤6:起草实验 → 获取反馈 → 修订
- [ ] 步骤7:起草相关工作 → 获取反馈 → 修订
- [ ] 步骤8:起草局限性 → 获取反馈 → 修订
- [ ] 步骤9:完成论文检查清单(必填)
- [ ] 步骤10:最终审核与提交步骤1:确定核心贡献(一句话)
此步骤需要科研人员的明确确认。
在撰写任何内容前,先明确并验证:
- 你的论文核心贡献是什么?
- 哪些内容是此前不存在或不明确的?
"我提议将核心贡献框架定为:'[一句话]'。这是否符合你认为的核心要点?是否需要调整侧重点?"
步骤2:起草图1
图1需要特别关注——许多读者会直接跳到图1。
- 传达核心思想、方法或最具说服力的结果
- 使用矢量图形(图表使用PDF/EPS格式)
- 撰写无需依赖正文即可独立理解的图注
- 确保黑白模式下的可读性(8%的男性有色觉障碍)
步骤3:撰写摘要(5句式公式)
源自Sebastian Farquhar(DeepMind):
1. 成果:"我们提出..."、"我们证明..."、"我们展示..."
2. 难度与重要性:说明为何这项工作困难且重要
3. 方法:使用专业关键词提升可发现性
4. 证据:支撑成果的证据
5. 核心数据:最显著的数值/结果删除诸如"大语言模型已取得显著成功..."这类通用开头。
步骤4:撰写引言(最多1-1.5页)
必须包含:
- 2-4条核心贡献要点(双栏格式下每条最多1-2行)
- 清晰的问题陈述
- 方法概述
- 方法部分需在第2-3页前开始
步骤5:方法章节
确保可复现性:
- 概念性大纲或伪代码
- 列出所有超参数
- 提供足够的架构细节以支持复现
- 呈现最终设计决策;消融实验放在实验部分
步骤6:实验章节
每个实验需明确说明:
- 它支撑哪个论点
- 与核心贡献的关联
- 实验设置(细节放在附录)
- 观察重点:"蓝线显示X,证明了Y"
要求:
- 误差棒及计算方法(标准差 vs 标准误)
- 超参数搜索范围
- 计算基础设施(GPU类型、总时长)
- 随机种子设置方法
步骤7:相关工作
按方法分类组织,而非按论文逐条罗列:
好的写法:"一类研究采用Floogledoodle的假设[引用],而我们采用Doobersnoddle的假设,因为..."
差的写法:"Snap等人提出X,而Crackle等人提出Y。"
慷慨引用——审稿人很可能是相关论文的作者。
步骤8:局限性章节(必填)
所有顶会都要求此章节。counter-intuitively的是,诚实反而有益:
- 审稿人被指示不会因诚实承认局限性而扣分
- 提前主动指出弱点,可预先化解批评
- 解释为何局限性不会削弱核心论点
步骤9:论文检查清单
NeurIPS、ICML和ICLR都要求提交论文检查清单。详见references/checklists.md。
Writing Philosophy for Top ML Conferences
顶会ML论文写作理念
This section distills the most important writing principles from leading ML researchers. These aren't optional style suggestions—they're what separates accepted papers from rejected ones.
"A paper is a short, rigorous, evidence-based technical story with a takeaway readers care about." — Neel Nanda
本节提炼了顶尖ML研究者最重要的写作原则。这些不是可选的风格建议——它们是区分录用论文与拒稿的关键。
"论文是一个简短、严谨、基于证据的技术故事,包含读者关心的核心观点。" — Neel Nanda
The Sources Behind This Guidance
指导来源
This skill synthesizes writing philosophy from researchers who have published extensively at top venues:
| Source | Key Contribution | Link |
|---|---|---|
| Neel Nanda (Google DeepMind) | The Narrative Principle, What/Why/So What framework | How to Write ML Papers |
| Sebastian Farquhar (DeepMind) | 5-sentence abstract formula | How to Write ML Papers |
| Gopen & Swan | 7 principles of reader expectations | Science of Scientific Writing |
| Zachary Lipton | Word choice, eliminating hedging | Heuristics for Scientific Writing |
| Jacob Steinhardt (UC Berkeley) | Precision, consistent terminology | Writing Tips |
| Ethan Perez (Anthropic) | Micro-level clarity tips | Easy Paper Writing Tips |
| Andrej Karpathy | Single contribution focus | Various lectures |
For deeper dives into any of these, see:
- references/writing-guide.md - Full explanations with examples
- references/sources.md - Complete bibliography
本技能综合了在顶会发表大量论文的研究者的写作理念:
| 来源 | 核心贡献 | 链接 |
|---|---|---|
| Neel Nanda(Google DeepMind) | 叙事原则、What/Why/So What框架 | How to Write ML Papers |
| Sebastian Farquhar(DeepMind) | 5句式摘要公式 | How to Write ML Papers |
| Gopen & Swan | 7项读者预期原则 | Science of Scientific Writing |
| Zachary Lipton | 措辞选择、消除模糊表述 | Heuristics for Scientific Writing |
| Jacob Steinhardt(UC Berkeley) | 精确性、术语一致性 | Writing Tips |
| Ethan Perez(Anthropic) | 微观层面清晰度技巧 | Easy Paper Writing Tips |
| Andrej Karpathy | 聚焦单一核心贡献 | 各类讲座 |
如需深入了解,可参考:
- references/writing-guide.md - 完整解释及示例
- references/sources.md - 完整参考文献
Time Allocation (From Neel Nanda)
时间分配(来自Neel Nanda)
Spend approximately equal time on each of:
- The abstract
- The introduction
- The figures
- Everything else combined
Why? Most reviewers form judgments before reaching your methods. Readers encounter your paper as: title → abstract → introduction → figures → maybe the rest.
将大约相等的时间分配给以下四部分:
- 摘要
- 引言
- 图表
- 其余所有内容
原因? 大多数审稿人在看到方法部分前就已形成判断。读者接触论文的顺序是:标题 → 摘要 → 引言 → 图表 → 可能才会看其余内容。
Writing Style Guidelines
写作风格指南
Sentence-Level Clarity (Gopen & Swan's 7 Principles)
句子层面清晰度(Gopen & Swan的7项原则)
These principles are based on how readers actually process prose. Violating them forces readers to spend cognitive effort on structure rather than content.
| Principle | Rule | Example |
|---|---|---|
| Subject-verb proximity | Keep subject and verb close | ❌ "The model, which was trained on..., achieves" → ✅ "The model achieves... after training on..." |
| Stress position | Place emphasis at sentence ends | ❌ "Accuracy improves by 15% when using attention" → ✅ "When using attention, accuracy improves by 15%" |
| Topic position | Put context first, new info after | ✅ "Given these constraints, we propose..." |
| Old before new | Familiar info → unfamiliar info | Link backward, then introduce new |
| One unit, one function | Each paragraph makes one point | Split multi-point paragraphs |
| Action in verb | Use verbs, not nominalizations | ❌ "We performed an analysis" → ✅ "We analyzed" |
| Context before new | Set stage before presenting | Explain before showing equation |
Full 7 principles with detailed examples: See references/writing-guide.md
这些原则基于读者实际处理文本的方式。违反这些原则会让读者将认知精力浪费在结构上,而非内容上。
| 原则 | 规则 | 示例 |
|---|---|---|
| 主谓靠近 | 主语和谓语尽量靠近 | ❌ "经过...训练的模型,实现了" → ✅ "模型在经过...训练后实现了" |
| 强调位置 | 将重点放在句末 | ❌ "使用注意力机制时准确率提升15%" → ✅ "使用注意力机制时,准确率提升15%" |
| 主题位置 | 先给出上下文,再呈现新信息 | ✅ "基于这些约束,我们提出..." |
| 先旧后新 | 先呈现熟悉信息,再引入新信息 | 先回顾前文,再介绍新内容 |
| 一段一意 | 每个段落只表达一个核心观点 | 拆分包含多个观点的段落 |
| 动词表动作 | 使用动词,而非名词化形式 | ❌ "我们进行了分析" → ✅ "我们分析了" |
| 先上下文后新内容 | 先铺垫背景,再呈现内容 | 先解释,再展示公式 |
完整的7项原则及详细示例:详见references/writing-guide.md
Micro-Level Tips (Ethan Perez)
微观层面技巧(Ethan Perez)
These small changes accumulate into significantly clearer prose:
- Minimize pronouns: ❌ "This shows..." → ✅ "This result shows..."
- Verbs early: Position verbs near sentence start
- Unfold apostrophes: ❌ "X's Y" → ✅ "The Y of X" (when awkward)
- Delete filler words: "actually," "a bit," "very," "really," "basically," "quite," "essentially"
Full micro-tips with examples: See references/writing-guide.md
这些小改动累积起来能显著提升文本清晰度:
- 减少代词使用:❌ "这表明..." → ✅ "这一结果表明..."
- 动词前置:将动词放在句子靠前位置
- 展开所有格:❌ "X的Y" → ✅ "X的Y"(当表述生硬时调整)
- 删除填充词:"actually"、"a bit"、"very"、"really"、"basically"、"quite"、"essentially"
完整的微观技巧及示例:详见references/writing-guide.md
Word Choice (Zachary Lipton)
措辞选择(Zachary Lipton)
- Be specific: ❌ "performance" → ✅ "accuracy" or "latency" (say what you mean)
- Eliminate hedging: Drop "may" and "can" unless genuinely uncertain
- Avoid incremental vocabulary: ❌ "combine," "modify," "expand" → ✅ "develop," "propose," "introduce"
- Delete intensifiers: ❌ "provides very tight approximation" → ✅ "provides tight approximation"
- 具体明确:❌ "performance" → ✅ "accuracy"或"latency"(明确你的意思)
- 消除模糊表述:除非确实不确定,否则删除"may"和"can"
- 避免增量词汇:❌ "combine"、"modify"、"expand" → ✅ "develop"、"propose"、"introduce"
- 删除强化词:❌ "提供非常精确的近似" → ✅ "提供精确的近似"
Precision Over Brevity (Jacob Steinhardt)
精确优先于简洁(Jacob Steinhardt)
- Consistent terminology: Different terms for same concept creates confusion. Pick one and stick with it.
- State assumptions formally: Before theorems, list all assumptions explicitly
- Intuition + rigor: Provide intuitive explanations alongside formal proofs
- 术语一致:同一概念使用不同术语会造成混淆。选定一个术语并坚持使用。
- 明确陈述假设:在定理前,明确列出所有假设
- 直觉+严谨:在形式化证明旁提供直观解释
What Reviewers Actually Read
审稿人实际阅读内容
Understanding reviewer behavior helps prioritize your effort:
| Paper Section | % Reviewers Who Read | Implication |
|---|---|---|
| Abstract | 100% | Must be perfect |
| Introduction | 90%+ (skimmed) | Front-load contribution |
| Figures | Examined before methods | Figure 1 is critical |
| Methods | Only if interested | Don't bury the lede |
| Appendix | Rarely | Put only supplementary details |
Bottom line: If your abstract and intro don't hook reviewers, they may never read your brilliant methods section.
了解审稿人行为有助于优化精力分配:
| 论文章节 | 阅读的审稿人比例 | 启示 |
|---|---|---|
| 摘要 | 100% | 必须完美 |
| 引言 | 90%+(略读) | 前置核心贡献 |
| 图表 | 在方法部分前查看 | 图1至关重要 |
| 方法 | 仅在感兴趣时阅读 | 不要隐藏核心亮点 |
| 附录 | 很少阅读 | 仅放置补充细节 |
底线:如果你的摘要和引言无法吸引审稿人,他们可能永远不会读到你出色的方法部分。
Conference Requirements Quick Reference
会议要求速查
| Conference | Page Limit | Extra for Camera-Ready | Key Requirement |
|---|---|---|---|
| NeurIPS 2025 | 9 pages | +0 | Mandatory checklist, lay summary for accepted |
| ICML 2026 | 8 pages | +1 | Broader Impact Statement required |
| ICLR 2026 | 9 pages | +1 | LLM disclosure required, reciprocal reviewing |
| ACL 2025 | 8 pages (long) | varies | Limitations section mandatory |
| AAAI 2026 | 7 pages | +1 | Strict style file adherence |
| COLM 2025 | 9 pages | +1 | Focus on language models |
Universal Requirements:
- Double-blind review (anonymize submissions)
- References don't count toward page limit
- Appendices unlimited but reviewers not required to read
- LaTeX required for all venues
LaTeX Templates: See templates/ directory for all conference templates.
| 会议 | 页数限制 | 终稿额外页数 | 关键要求 |
|---|---|---|---|
| NeurIPS 2025 | 9页 | +0 | 强制检查清单,录用后需提供通俗摘要 |
| ICML 2026 | 8页 | +1 | 必须包含广泛影响声明 |
| ICLR 2026 | 9页 | +1 | 必须披露LLM使用情况,需同意互评协议 |
| ACL 2025 | 8页(长文) | 依情况而定 | 强制包含局限性章节 |
| AAAI 2026 | 7页 | +1 | 严格遵循格式文件 |
| COLM 2025 | 9页 | +1 | 聚焦大语言模型 |
通用要求:
- 双盲评审(提交时需匿名)
- 参考文献不计入页数限制
- 附录页数不限,但审稿人无义务阅读
- 所有会议均要求使用LaTeX
LaTeX模板:详见templates/目录下的各会议模板。
Using LaTeX Templates Properly
正确使用LaTeX模板
Workflow 4: Starting a New Paper from Template
工作流4:从模板开始撰写新论文
Always copy the entire template directory first, then write within it.
Template Setup Checklist:
- [ ] Step 1: Copy entire template directory to new project
- [ ] Step 2: Verify template compiles as-is (before any changes)
- [ ] Step 3: Read the template's example content to understand structure
- [ ] Step 4: Replace example content section by section
- [ ] Step 5: Keep template comments/examples as reference until done
- [ ] Step 6: Clean up template artifacts only at the endStep 1: Copy the Full Template
bash
undefined始终先复制完整的模板目录,再在其中撰写内容。
模板设置清单:
- [ ] 步骤1:将完整模板目录复制到新项目
- [ ] 步骤2:验证模板可直接编译(修改前)
- [ ] 步骤3:阅读模板示例内容以了解结构
- [ ] 步骤4:逐节替换示例内容
- [ ] 步骤5:保留模板注释/示例作为参考直至完成
- [ ] 步骤6:仅在最后清理模板遗留内容步骤1:复制完整模板
bash
undefinedCreate your paper directory with the complete template
创建包含完整模板的论文目录
cp -r templates/neurips2025/ ~/papers/my-new-paper/
cd ~/papers/my-new-paper/
cp -r templates/neurips2025/ ~/papers/my-new-paper/
cd ~/papers/my-new-paper/
Verify structure is complete
验证结构完整
ls -la
ls -la
Should see: main.tex, neurips.sty, Makefile, etc.
应包含:main.tex、neurips.sty、Makefile等
**⚠️ IMPORTANT**: Copy the ENTIRE directory, not just `main.tex`. Templates include:
- Style files (`.sty`) - required for compilation
- Bibliography styles (`.bst`) - required for references
- Example content - useful as reference
- Makefiles - for easy compilation
**Step 2: Verify Template Compiles First**
Before making ANY changes, compile the template as-is:
```bash
**⚠️ 重要提示**:复制**整个目录**,而非仅`main.tex`。模板包含:
- 样式文件(`.sty`) - 编译必需
- 参考文献样式(`.bst`) - 引用格式必需
- 示例内容 - 可作为参考
- Makefile - 简化编译流程
**步骤2:先验证模板可编译**
在进行**任何**修改前,先直接编译模板:
```bashUsing latexmk (recommended)
使用latexmk(推荐)
latexmk -pdf main.tex
latexmk -pdf main.tex
Or manual compilation
或手动编译
pdflatex main.tex
bibtex main
pdflatex main.tex
pdflatex main.tex
If the unmodified template doesn't compile, fix that first. Common issues:
- Missing TeX packages → install via `tlmgr install <package>`
- Wrong TeX distribution → use TeX Live (recommended)
**Step 3: Keep Template Content as Reference**
Don't immediately delete all example content. Instead:
```latex
% KEEP template examples commented out as you write
% This shows you the expected format
% Template example (keep for reference):
% \begin{figure}[t]
% \centering
% \includegraphics[width=0.8\linewidth]{example-image}
% \caption{Template shows caption style}
% \end{figure}
% Your actual figure:
\begin{figure}[t]
\centering
\includegraphics[width=0.8\linewidth]{your-figure.pdf}
\caption{Your caption following the same style.}
\end{figure}Step 4: Replace Content Section by Section
Work through the paper systematically:
Replacement Order:
1. Title and authors (anonymize for submission)
2. Abstract
3. Introduction
4. Methods
5. Experiments
6. Related Work
7. Conclusion
8. References (your .bib file)
9. AppendixFor each section:
- Read the template's example content
- Note any special formatting or macros used
- Replace with your content following the same patterns
- Compile frequently to catch errors early
Step 5: Use Template Macros
Templates often define useful macros. Check the preamble for:
latex
% Common template macros to use:
\newcommand{\method}{YourMethodName} % Consistent method naming
\newcommand{\eg}{e.g.,\xspace} % Proper abbreviations
\newcommand{\ie}{i.e.,\xspace}
\newcommand{\etal}{\textit{et al.}\xspace}Step 6: Clean Up Only at the End
Only remove template artifacts when paper is nearly complete:
latex
% BEFORE SUBMISSION - remove these:
% - Commented-out template examples
% - Unused packages
% - Template's example figures/tables
% - Lorem ipsum or placeholder text
% KEEP these:
% - All style files (.sty)
% - Bibliography style (.bst)
% - Required packages from template
% - Any custom macros you're usingpdflatex main.tex
bibtex main
pdflatex main.tex
pdflatex main.tex
如果未修改的模板无法编译,先解决此问题。常见问题:
- 缺少TeX包 → 通过`tlmgr install <package>`安装
- TeX发行版错误 → 使用TeX Live(推荐)
**步骤3:保留模板内容作为参考**
不要立即删除所有示例内容。相反:
```latex
% 保留模板示例并注释,作为写作参考
% 这能让你了解预期格式
% 模板示例(保留参考):
% \begin{figure}[t]
% \centering
% \includegraphics[width=0.8\linewidth]{example-image}
% \caption{模板展示了图注样式}
% \end{figure}
% 你的实际图表:
\begin{figure}[t]
\centering
\includegraphics[width=0.8\linewidth]{your-figure.pdf}
\caption{你的图注遵循相同样式。}
\end{figure}步骤4:逐节替换内容
系统地完成论文各部分:
替换顺序:
1. 标题和作者(提交时匿名)
2. 摘要
3. 引言
4. 方法
5. 实验
6. 相关工作
7. 结论
8. 参考文献(你的.bib文件)
9. 附录对于每个章节:
- 阅读模板的示例内容
- 注意使用的特殊格式或宏
- 替换为你的内容并遵循相同模式
- 频繁编译以尽早发现错误
步骤5:使用模板宏
模板通常定义了实用的宏。查看导言区:
latex
% 可使用的常见模板宏:
\newcommand{\method}{YourMethodName} % 方法名称保持一致
\newcommand{\eg}{e.g.,\xspace} % 正确的缩写
\newcommand{\ie}{i.e.,\xspace}
\newcommand{\etal}{\textit{et al.}\xspace}步骤6:仅在最后清理
仅在论文接近完成时才移除模板遗留内容:
latex
% 提交前 - 删除以下内容:
% - 注释掉的模板示例
% - 未使用的包
% - 模板的示例图表/表格
% - 占位文本
% 保留以下内容:
% - 所有样式文件(.sty)
% - 参考文献样式(.bst)
% - 模板要求的必需包
% - 你使用的自定义宏Template Pitfalls to Avoid
需避免的模板陷阱
| Pitfall | Problem | Solution |
|---|---|---|
Copying only | Missing | Copy entire directory |
Modifying | Breaks conference formatting | Never edit style files |
| Adding random packages | Conflicts, breaks template | Only add if necessary |
| Deleting template content too early | Lose formatting reference | Keep as comments until done |
| Not compiling frequently | Errors accumulate | Compile after each section |
| 陷阱 | 问题 | 解决方案 |
|---|---|---|
仅复制 | 缺少 | 复制整个目录 |
修改 | 破坏会议格式 | 绝不编辑样式文件 |
| 随意添加包 | 冲突,破坏模板 | 仅在必要时添加 |
| 过早删除模板内容 | 失去格式参考 | 保留注释直至完成 |
| 不频繁编译 | 错误累积 | 每完成一个章节就编译一次 |
Quick Template Reference
模板速查
| Conference | Main File | Key Style File | Notes |
|---|---|---|---|
| NeurIPS 2025 | | | Has Makefile |
| ICML 2026 | | | Includes algorithm packages |
| ICLR 2026 | | | Has math_commands.tex |
| ACL | | | Strict formatting |
| AAAI 2026 | | | Very strict compliance |
| COLM 2025 | | | Similar to ICLR |
| 会议 | 主文件 | 关键样式文件 | 说明 |
|---|---|---|---|
| NeurIPS 2025 | | | 包含Makefile |
| ICML 2026 | | | 包含算法包 |
| ICLR 2026 | | | 包含math_commands.tex |
| ACL | | | 格式严格 |
| AAAI 2026 | | | 合规要求极高 |
| COLM 2025 | | | 与ICLR类似 |
Conference Resubmission & Format Conversion
会议转投与格式转换
When a paper is rejected or withdrawn from one venue and resubmitted to another, format conversion is required. This is a common workflow in ML research.
当论文被拒稿或从一个会议撤回并转投至另一个会议时,需要进行格式转换。这是ML研究中的常见工作流。
Workflow 3: Converting Between Conference Formats
工作流3:会议格式转换
Format Conversion Checklist:
- [ ] Step 1: Identify source and target template differences
- [ ] Step 2: Create new project with target template
- [ ] Step 3: Copy content sections (not preamble)
- [ ] Step 4: Adjust page limits and content
- [ ] Step 5: Update conference-specific requirements
- [ ] Step 6: Verify compilation and formattingStep 1: Key Template Differences
| From → To | Page Change | Key Adjustments |
|---|---|---|
| NeurIPS → ICML | 9 → 8 pages | Cut 1 page, add Broader Impact if missing |
| ICML → ICLR | 8 → 9 pages | Can expand experiments, add LLM disclosure |
| NeurIPS → ACL | 9 → 8 pages | Restructure for NLP conventions, add Limitations |
| ICLR → AAAI | 9 → 7 pages | Significant cuts needed, strict style adherence |
| Any → COLM | varies → 9 | Reframe for language model focus |
Step 2: Content Migration (NOT Template Merge)
Never copy LaTeX preambles between templates. Instead:
bash
undefined格式转换清单:
- [ ] 步骤1:识别源模板与目标模板的差异
- [ ] 步骤2:使用目标模板创建新项目
- [ ] 步骤3:复制内容章节(而非导言区)
- [ ] 步骤4:调整页数限制和内容
- [ ] 步骤5:更新会议特定要求
- [ ] 步骤6:验证编译和格式步骤1:模板关键差异
| 源会议 → 目标会议 | 页数变化 | 核心调整 |
|---|---|---|
| NeurIPS → ICML | 9 → 8页 | 缩减1页,添加缺失的广泛影响声明 |
| ICML → ICLR | 8 → 9页 | 可扩充实验,添加LLM披露 |
| NeurIPS → ACL | 9 → 8页 | 按NLP惯例重构,添加局限性章节 |
| ICLR → AAAI | 9 → 7页 | 大幅精简,严格遵循格式 |
| 任意 → COLM | 可变 → 9页 | 调整为聚焦大语言模型 |
步骤2:内容迁移(而非模板合并)
绝不要在模板间复制LaTeX导言区。 正确做法:
bash
undefined1. Start fresh with target template
1. 从目标模板开始
cp -r templates/icml2026/ new_submission/
cp -r templates/icml2026/ new_submission/
2. Copy ONLY content sections from old paper
2. 仅复制旧论文的内容章节
- Abstract text
- 摘要文本
- Section content (between \section{} commands)
- 章节内容(在\section{}命令之间)
- Figures and tables
- 图表
- Bibliography entries
- 参考文献条目
3. Paste into target template structure
3. 粘贴到目标模板结构中
**Step 3: Adjusting for Page Limits**
When cutting pages (e.g., NeurIPS 9 → AAAI 7):
- Move detailed proofs to appendix
- Condense related work (cite surveys instead of individual papers)
- Combine similar experiments into unified tables
- Use smaller figure sizes with subfigures
- Tighten writing: eliminate redundancy, use active voice
When expanding (e.g., ICML 8 → ICLR 9):
- Add ablation studies reviewers requested
- Expand limitations discussion
- Include additional baselines
- Add qualitative examples
**Step 4: Conference-Specific Adjustments**
| Target Venue | Required Additions |
|--------------|-------------------|
| **ICML** | Broader Impact Statement (after conclusion) |
| **ICLR** | LLM usage disclosure, reciprocal reviewing agreement |
| **ACL/EMNLP** | Limitations section (mandatory), Ethics Statement |
| **AAAI** | Strict adherence to style file (no modifications) |
| **NeurIPS** | Paper checklist (appendix), lay summary if accepted |
**Step 5: Update References**
```latex
% Remove self-citations that reveal identity (for blind review)
% Update any "under review" citations to published versions
% Add new relevant work published since last submissionStep 6: Addressing Previous Reviews
When resubmitting after rejection:
- Do address reviewer concerns in the new version
- Do add experiments/clarifications reviewers requested
- Don't include a "changes from previous submission" section (blind review)
- Don't reference the previous submission or reviews
Common Conversion Pitfalls:
- ❌ Copying commands (causes conflicts)
\usepackage - ❌ Keeping old conference header/footer commands
- ❌ Forgetting to update path
\bibliography{} - ❌ Missing conference-specific required sections
- ❌ Exceeding page limit after format change
**步骤3:适配页数限制**
当需要缩减页数时(如NeurIPS 9页 → AAAI 7页):
- 将详细证明移至附录
- 精简相关工作(引用综述而非单篇论文)
- 将相似实验合并为统一表格
- 使用更小的图表尺寸和子图
- 精简写作:消除冗余,使用主动语态
当可以扩充页数时(如ICML 8页 → ICLR 9页):
- 添加审稿人要求的消融研究
- 扩充局限性讨论
- 包含更多基线方法
- 添加定性示例
**步骤4:会议特定调整**
| 目标会议 | 必需添加内容 |
|--------------|-------------------|
| **ICML** | 广泛影响声明(结论后) |
| **ICLR** | LLM使用披露,互评协议同意声明 |
| **ACL/EMNLP** | 强制局限性章节,伦理声明 |
| **AAAI** | 严格遵循样式文件(禁止修改) |
| **NeurIPS** | 论文检查清单(附录),录用后提供通俗摘要 |
**步骤5:更新参考文献**
```latex
% 删除可能暴露身份的自引(双盲评审)
% 将任何"审稿中"的引用更新为已发表版本
% 添加上次提交后发表的相关新研究步骤6:回应此前的审稿意见
拒稿后转投时:
- 务必在新版本中回应审稿人关注点
- 务必添加审稿人要求的实验/澄清内容
- 不要包含"与上次提交的变化"章节(双盲评审)
- 不要提及此前的提交或审稿意见
常见转换陷阱:
- ❌ 复制命令(导致冲突)
\usepackage - ❌ 保留旧会议的页眉/页脚命令
- ❌ 忘记更新路径
\bibliography{} - ❌ 遗漏会议特定的必填章节
- ❌ 格式转换后超出页数限制
Citation Workflow (Hallucination Prevention)
引用工作流(防止虚构)
⚠️ CRITICAL: AI-generated citations have ~40% error rate. Never write BibTeX from memory.
⚠️ 重中之重:AI生成的引用错误率约为40%。绝不能凭记忆编写BibTeX。
The Golden Rule
黄金法则
IF you cannot verify a citation through web search:
→ Mark it as [CITATION NEEDED] or [PLACEHOLDER - VERIFY]
→ Tell the scientist explicitly
→ NEVER invent a plausible-sounding referenceMANDATORY: Use WebSearch tool to verify EVERY citation before adding to bibliography.
如果无法通过网页搜索验证引用:
→ 标记为[CITATION NEEDED]或[PLACEHOLDER - VERIFY]
→ 明确告知科研人员
→ 绝不要编造看似合理的参考文献强制要求:在将引用添加到参考文献前,必须使用WebSearch工具每一条都验证。
Workflow 2: Adding Citations
工作流2:添加引用
Citation Verification (MANDATORY for every citation):
- [ ] Step 1: Use WebSearch to find the paper
- [ ] Step 2: Verify paper exists on Google Scholar
- [ ] Step 3: Confirm paper details (title, authors, year, venue)
- [ ] Step 4: Retrieve BibTeX from Google Scholar or DOI
- [ ] Step 5: Verify the claim you're citing actually appears in the paper
- [ ] Step 6: Add verified BibTeX to bibliography
- [ ] Step 7: If ANY step fails → mark as placeholder, inform scientistStep 1: Use WebSearch to Find the Paper
When you need to cite a paper, ALWAYS start with web search:
WebSearch query examples:
- "Attention is All You Need Vaswani 2017"
- "RLHF language model alignment 2023"
- "sparse autoencoders interpretability Anthropic"
- "transformer architecture NeurIPS"What to look for in search results:
- Paper title matches your intended citation
- Authors are correct
- Publication year is correct
- Venue (conference/journal) is identified
Step 2: Verify on Google Scholar
After finding the paper, verify it exists on Google Scholar:
WebSearch query: "site:scholar.google.com [paper title] [first author]"
Example: "site:scholar.google.com Attention is All You Need Vaswani"Verification checklist:
- ✅ Paper appears in Google Scholar results
- ✅ Title matches exactly (or very close)
- ✅ Authors match
- ✅ Year matches
- ✅ Venue is listed (conference/journal)
- ✅ Citation count is reasonable (not 0 for old papers)
If paper NOT found on Google Scholar:
- ❌ STOP - Do not cite
- Mark as
[CITATION NEEDED - not found on Google Scholar] - Inform scientist explicitly
Step 3: Confirm Paper Details
Before retrieving BibTeX, double-check all details:
Verification checklist:
- Title: [exact title from Google Scholar]
- Authors: [all authors, in order]
- Year: [publication year]
- Venue: [conference/journal name]
- DOI: [if available]Step 4: Retrieve BibTeX
Option 1: From Google Scholar (Recommended)
- Find the paper on Google Scholar
- Click "Cite" button below the paper
- Select "BibTeX" format
- Copy the BibTeX entry
Option 2: From DOI (if available)
- Use WebSearch to find:
"doi.org/[DOI]" - Look for BibTeX export option on the publisher's page
- Copy the BibTeX entry
Option 3: From arXiv (for preprints)
- Find paper on arXiv
- Click "Export BibTeX Citation" on the right sidebar
- Copy the BibTeX entry
CRITICAL: Never write BibTeX from memory. Always copy from verified source.
Step 5: Verify the Claim
Before citing for a specific claim, verify the claim actually appears in the paper:
Verification process:
1. Use WebSearch to access the paper (PDF or HTML)
2. Search for keywords related to your claim
3. Confirm the claim is explicitly stated or clearly implied
4. Note the section/page where claim appearsIf you cannot access the paper:
- ❌ Do not cite for specific claims
- Only cite for general contributions (if verified on Google Scholar)
- Mark as
[CLAIM NOT VERIFIED - no access to paper]
Step 6: Add Verified BibTeX to Bibliography
Only after completing all verification steps:
latex
% Add to your .bib file
@inproceedings{vaswani2017attention,
title={Attention is All You Need},
author={Vaswani, Ashish and Shazeer, Noam and ...},
booktitle={Advances in Neural Information Processing Systems},
year={2017}
}
% Use in your paper
\cite{vaswani2017attention}Step 7: Handle Failures Explicitly
If you cannot verify a citation at ANY step:
latex
% Option 1: Explicit placeholder
\cite{PLACEHOLDER_smith2023_verify} % TODO: Could not verify - scientist must confirm
% Option 2: Note in text
... as shown in prior work [CITATION NEEDED - could not verify Smith et al. 2023].Always inform the scientist:
"I could not verify the following citations and have marked them as placeholders:
- Smith et al. 2023 on reward hacking - not found on Google Scholar
- Jones 2022 on scaling laws - found similar paper but different authors Please verify these before submission."
引用验证(每条引用都必须执行):
- [ ] 步骤1:使用WebSearch查找论文
- [ ] 步骤2:在Google Scholar上验证论文存在
- [ ] 步骤3:确认论文细节(标题、作者、年份、会议)
- [ ] 步骤4:从Google Scholar或DOI获取BibTeX
- [ ] 步骤5:验证你引用的论点确实出现在论文中
- [ ] 步骤6:将验证后的BibTeX添加到参考文献
- [ ] 步骤7:若任何步骤失败 → 标记为占位符,告知科研人员步骤1:使用WebSearch查找论文
当需要引用某篇论文时,始终从网页搜索开始:
WebSearch查询示例:
- "Attention is All You Need Vaswani 2017"
- "RLHF language model alignment 2023"
- "sparse autoencoders interpretability Anthropic"
- "transformer architecture NeurIPS"搜索结果关注点:
- 论文标题与预期引用匹配
- 作者正确
- 年份正确
- 会议/期刊明确
步骤2:在Google Scholar上验证
找到论文后,在Google Scholar上验证其存在:
WebSearch查询:"site:scholar.google.com [论文标题] [第一作者]"
示例:"site:scholar.google.com Attention is All You Need Vaswani"验证清单:
- ✅ 论文出现在Google Scholar结果中
- ✅ 标题完全匹配(或高度相似)
- ✅ 作者匹配
- ✅ 年份匹配
- ✅ 会议/期刊已列出
- ✅ 引用量合理(旧论文引用量不为0)
如果在Google Scholar上找不到论文:
- ❌ 停止 - 不要引用
- 标记为
[CITATION NEEDED - not found on Google Scholar] - 明确告知科研人员
步骤3:确认论文细节
获取BibTeX前,再次核对所有细节:
验证清单:
- 标题:[Google Scholar上的精确标题]
- 作者:[所有作者,按顺序]
- 年份:[发表年份]
- 会议:[会议/期刊名称]
- DOI:[若有]步骤4:获取BibTeX
选项1:从Google Scholar获取(推荐)
- 在Google Scholar上找到论文
- 点击论文下方的"引用"按钮
- 选择"BibTeX"格式
- 复制BibTeX条目
选项2:从DOI获取(若有)
- 使用WebSearch查找:
"doi.org/[DOI]" - 在出版商页面查找BibTeX导出选项
- 复制BibTeX条目
选项3:从arXiv获取(预印本)
- 在arXiv上找到论文
- 点击右侧边栏的"Export BibTeX Citation"
- 复制BibTeX条目
关键提示:绝不能凭记忆生成BibTeX。始终从已验证的来源复制。
步骤5:验证论点
为特定论点引用前,验证该论点确实出现在论文中:
验证流程:
1. 通过WebSearch获取论文(PDF或HTML)
2. 搜索与你的论点相关的关键词
3. 确认论点被明确陈述或清晰暗示
4. 记录论点所在的章节/页码如果无法获取论文:
- ❌ 不要为特定论点引用
- 仅可引用其通用贡献(若已在Google Scholar上验证)
- 标记为
[CLAIM NOT VERIFIED - no access to paper]
步骤6:将验证后的BibTeX添加到参考文献
仅在完成所有验证步骤后:
latex
% 添加到你的.bib文件
@inproceedings{vaswani2017attention,
title={Attention is All You Need},
author={Vaswani, Ashish and Shazeer, Noam and ...},
booktitle={Advances in Neural Information Processing Systems},
year={2017}
}
% 在论文中引用
\cite{vaswani2017attention}步骤7:处理验证失败的情况
如果任何步骤无法验证引用:
latex
% 选项1:明确占位符
\cite{PLACEHOLDER_smith2023_verify} % TODO: 无法验证 - 请科研人员确认
% 选项2:在正文中标注
... 如先前研究所示[CITATION NEEDED - 无法验证Smith等人2023年的论文]。务必告知科研人员:
"我无法验证以下引用,已将其标记为占位符:
- Smith等人2023年关于奖励黑客的论文 - Google Scholar上未找到
- Jones2022年关于缩放定律的论文 - 找到相似论文但作者不同 提交前请验证这些引用。"
Summary: Citation Rules
总结:引用规则
| Situation | Action |
|---|---|
| Found on Google Scholar, verified details, got BibTeX | ✅ Use the citation |
| Found paper, verified on Google Scholar, no BibTeX | ✅ Create BibTeX from Google Scholar info |
| Paper exists but details don't match | ⚠️ Mark placeholder, inform scientist |
| Not found on Google Scholar | ❌ Mark |
| "I think there's a paper about X" | ❌ NEVER cite - search first or mark placeholder |
🚨 NEVER generate BibTeX from memory—always verify through WebSearch and Google Scholar. 🚨
| 情况 | 行动 |
|---|---|
| Google Scholar找到,细节验证通过,获取到BibTeX | ✅ 使用此引用 |
| 找到论文,Google Scholar验证通过,无BibTeX | ✅ 根据Google Scholar信息创建BibTeX |
| 论文存在但细节不匹配 | ⚠️ 标记占位符,告知科研人员 |
| Google Scholar上找不到 | ❌ 标记 |
| "我记得有一篇关于X的论文" | ❌ 绝不要引用 - 先搜索或标记占位符 |
🚨 绝不能凭记忆生成BibTeX——始终通过WebSearch和Google Scholar验证。 🚨
Complete Citation Workflow Example
完整引用工作流示例
Scenario: You need to cite the Transformer paper.
Step 1: WebSearch
Query: "Attention is All You Need Vaswani 2017"
Result: Found paper on multiple sources
Step 2: Google Scholar Verification
Query: "site:scholar.google.com Attention is All You Need Vaswani"
Result: ✅ Paper found, 50,000+ citations, NeurIPS 2017
Step 3: Confirm Details
- Title: "Attention is All You Need"
- Authors: Vaswani, Ashish; Shazeer, Noam; Parmar, Niki; ...
- Year: 2017
- Venue: NeurIPS (NIPS)
- DOI: Available
Step 4: Retrieve BibTeX
- Click "Cite" on Google Scholar
- Select BibTeX format
- Copy entry
Step 5: Verify Claim
- Access paper via WebSearch
- Confirm claim appears in paper
- Note section/page
Step 6: Add to Bibliography
- Paste BibTeX to .bib file
- Use \cite{vaswani2017attention} in paper
Step 7: Success
- Citation verified and added
- No placeholder needed场景:你需要引用Transformer论文。
步骤1:WebSearch
查询:"Attention is All You Need Vaswani 2017"
结果:在多个来源找到该论文
步骤2:Google Scholar验证
查询:"site:scholar.google.com Attention is All You Need Vaswani"
结果:✅ 找到论文,引用量50,000+,发表于NeurIPS 2017
步骤3:确认细节
- 标题:"Attention is All You Need"
- 作者:Vaswani, Ashish; Shazeer, Noam; Parmar, Niki; ...
- 年份:2017
- 会议:NeurIPS(原NIPS)
- DOI:可用
步骤4:获取BibTeX
- 点击Google Scholar上的"引用"
- 选择BibTeX格式
- 复制条目
步骤5:验证论点
- 通过WebSearch获取论文
- 确认论点出现在论文中
- 记录章节/页码
步骤6:添加到参考文献
- 将BibTeX粘贴到.bib文件
- 在论文中使用\cite{vaswani2017attention}
步骤7:完成
- 引用已验证并添加
- 无需占位符Common Issues and Solutions
常见问题与解决方案
Issue: Abstract too generic
Delete first sentence if it could be prepended to any ML paper. Start with your specific contribution.
Issue: Introduction exceeds 1.5 pages
Split background into Related Work. Front-load contribution bullets. Methods should start by page 2-3.
Issue: Experiments lack explicit claims
Add sentence before each experiment: "This experiment tests whether [specific claim]..."
Issue: Reviewers find paper hard to follow
- Add explicit signposting: "In this section, we show X"
- Use consistent terminology throughout
- Include figure captions that stand alone
Issue: Missing statistical significance
Always include:
- Error bars (specify: std dev or std error)
- Number of runs
- Statistical tests if comparing methods
问题:摘要过于通用
如果第一句可用于任何ML论文,直接删除。从你的具体贡献开始。
问题:引言超过1.5页
将背景内容拆分到相关工作部分。前置核心贡献要点。方法部分需在第2-3页前开始。
问题:实验缺乏明确论点
在每个实验前添加句子:"本实验测试[特定论点]是否成立..."
问题:审稿人认为论文难以理解
- 添加明确的引导语:"在本节中,我们展示X"
- 全程使用一致的术语
- 编写可独立理解的图注
问题:缺少统计显著性
始终包含:
- 误差棒(注明:标准差或标准误)
- 重复实验次数
- 对比方法时的统计检验
Reviewer Evaluation Criteria
审稿人评估标准
Reviewers assess papers on four dimensions:
| Criterion | What Reviewers Look For |
|---|---|
| Quality | Technical soundness, well-supported claims |
| Clarity | Clear writing, reproducible by experts |
| Significance | Community impact, advances understanding |
| Originality | New insights (doesn't require new method) |
Scoring (NeurIPS 6-point scale):
- 6: Strong Accept - Groundbreaking, flawless
- 5: Accept - Technically solid, high impact
- 4: Borderline Accept - Solid, limited evaluation
- 3: Borderline Reject - Solid but weaknesses outweigh
- 2: Reject - Technical flaws
- 1: Strong Reject - Known results or ethics issues
See references/reviewer-guidelines.md for detailed reviewer instructions.
审稿人从四个维度评估论文:
| 标准 | 审稿人关注点 |
|---|---|
| 质量 | 技术严谨性,论点支撑充分 |
| 清晰度 | 写作清晰,专家可复现 |
| 重要性 | 社区影响力,推动领域认知 |
| 原创性 | 新见解(无需全新方法) |
评分(NeurIPS 6分制):
- 6分:强烈录用 - 开创性,无瑕疵
- 5分:录用 - 技术扎实,高影响力
- 4分: borderline录用 - 扎实但评估有限
- 3分: borderline拒稿 - 扎实但弱点多于优势
- 2分:拒稿 - 技术缺陷
- 1分:强烈拒稿 - 已有成果或伦理问题
详细的审稿人指南请见references/reviewer-guidelines.md。
Tables and Figures
表格与图表
Tables
表格
Use LaTeX package for professional tables:
booktabslatex
\usepackage{booktabs}
\begin{tabular}{lcc}
\toprule
Method & Accuracy ↑ & Latency ↓ \\
\midrule
Baseline & 85.2 & 45ms \\
\textbf{Ours} & \textbf{92.1} & 38ms \\
\bottomrule
\end{tabular}Rules:
- Bold best value per metric
- Include direction symbols (↑ higher is better, ↓ lower is better)
- Right-align numerical columns
- Consistent decimal precision
使用LaTeX包制作专业表格:
booktabslatex
\usepackage{booktabs}
\begin{tabular}{lcc}
\toprule
Method & Accuracy ↑ & Latency ↓ \\
\midrule
Baseline & 85.2 & 45ms \\
\textbf{Ours} & \textbf{92.1} & 38ms \\
\bottomrule
\end{tabular}规则:
- 每个指标的最优值加粗
- 包含方向符号(↑ 越高越好,↓ 越低越好)
- 数值列右对齐
- 小数精度一致
Figures
图表
- Vector graphics (PDF, EPS) for all plots and diagrams
- Raster (PNG 600 DPI) only for photographs
- Use colorblind-safe palettes (Okabe-Ito or Paul Tol)
- Verify grayscale readability (8% of men have color vision deficiency)
- No title inside figure—the caption serves this function
- Self-contained captions—reader should understand without main text
- 矢量图形(PDF、EPS)用于所有图表和示意图
- 光栅图(PNG 600 DPI)仅用于照片
- 使用色觉障碍友好的配色方案(Okabe-Ito或Paul Tol)
- 验证灰度可读性(8%的男性有色觉障碍)
- 图表内无标题——图注承担此功能
- 可独立理解的图注——读者无需依赖正文即可理解
References & Resources
参考与资源
Reference Documents (Deep Dives)
参考文档(深度内容)
| Document | Contents |
|---|---|
| writing-guide.md | Gopen & Swan 7 principles, Ethan Perez micro-tips, word choice |
| citation-workflow.md | Citation APIs, Python code, BibTeX management |
| checklists.md | NeurIPS 16-item, ICML, ICLR, ACL requirements |
| reviewer-guidelines.md | Evaluation criteria, scoring, rebuttals |
| sources.md | Complete bibliography of all sources |
| Literature Research: | |
| arxiv-search-guide.md | arXiv search strategies, URL patterns, Chrome MCP automation |
| paper-quality-criteria.md | 5-dimension paper evaluation rubrics (innovation, method, experiments, writing, impact) |
| 文档 | 内容 |
|---|---|
| writing-guide.md | Gopen & Swan的7项原则、Ethan Perez的微观技巧、措辞选择 |
| citation-workflow.md | 引用API、Python代码、BibTeX管理 |
| checklists.md | NeurIPS 16项清单、ICML、ICLR、ACL要求 |
| reviewer-guidelines.md | 评估标准、评分、反驳策略 |
| sources.md | 所有来源的完整参考文献 |
| 文献调研: | |
| arxiv-search-guide.md | arXiv搜索策略、URL模式、Chrome MCP自动化 |
| paper-quality-criteria.md | 5维度论文评估细则(创新性、方法、实验、写作、影响力) |
LaTeX Templates
LaTeX模板
Templates in directory: ICML 2026, ICLR 2026, NeurIPS 2025, ACL/EMNLP, AAAI 2026, COLM 2025.
templates/Compiling to PDF:
- VS Code/Cursor: Install LaTeX Workshop extension + TeX Live → Save to auto-compile
- Command line: or
latexmk -pdf main.tex+pdflatexworkflowbibtex - Online: Upload to Overleaf
See templates/README.md for detailed setup instructions.
templates/编译为PDF:
- VS Code/Cursor:安装LaTeX Workshop扩展 + TeX Live → 保存时自动编译
- 命令行:或
latexmk -pdf main.tex+pdflatexbibtex工作流 - 在线工具:上传至Overleaf
详细的设置说明请见templates/README.md。
Key External Sources
核心外部资源
Writing Philosophy:
- Neel Nanda: How to Write ML Papers - Narrative, "What/Why/So What"
- Farquhar: How to Write ML Papers - 5-sentence abstract
- Gopen & Swan: Science of Scientific Writing - 7 reader expectation principles
- Lipton: Heuristics for Scientific Writing - Word choice
- Perez: Easy Paper Writing Tips - Micro-level clarity
写作理念:
- Neel Nanda: How to Write ML Papers - 叙事、"What/Why/So What"
- Farquhar: How to Write ML Papers - 5句式摘要
- Gopen & Swan: Science of Scientific Writing - 7项读者预期原则
- Lipton: Heuristics for Scientific Writing - 措辞选择
- Perez: Easy Paper Writing Tips - 微观层面清晰度