research-paper-writing

Guides end-to-end ML research paper production from experiment design through conference submission.

Updated May 4, 2026
One-click install
npx skills add https://github.com/Junkfooooood/hermes-governance --skill research-paper-writing-junkfooooood
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: research-paper-writing
Source: https://github.com/Junkfooooood/hermes-governance/tree/main/skills/research/research-paper-writing
Command: npx skills add https://github.com/Junkfooooood/hermes-governance --skill research-paper-writing-junkfooooood

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires semanticscholar, arxiv, habanero, requests, scipy, numpy, matplotlib, SciencePlots, and includes references (resource) components.

What problem does it solve? Writing a publication-ready ML/AI paper involves coordinating literature review, experiment design, statistical analysis, LaTeX drafting, citation verification, and venue-specific formatting — a process where missed steps (hallucinated citations, missing checklists, weak baselines) cause desk rejections. ## Core Features & Use Cases - Full Research Lifecycle: Eight iterative phases covering project setup, literature review, experiment design, execution monitoring, statistical analysis, drafting, self-review, and submission for NeurIPS, ICML, ICLR, ACL, AAAI, and COLM. - Citation Verification Workflow: Programmatic BibTeX retrieval via Semantic Scholar, CrossRef, and arXiv APIs with a mandatory 5-step verification process to prevent hallucinated references. - Venue Templates & Checklists: Official LaTeX templates for six conferences plus pre-submission checklists covering page limits, reproducibility statements, and ethics requirements. - Use Case: A researcher with experimental results asks the agent to draft an ICML submission — the skill produces a grounded draft from the experiment log, verifies every citation via DOI content negotiation, applies the ICML 2026 template, and runs a simulated reviewer pass before submission. ## Quick Start Use the research-paper-writing skill to draft a NeurIPS paper from the experiment results in my results/ directory, with verified citations and the official template.

Frequently Asked Questions about research-paper-writing

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I write a machine learning paper for NeurIPS or ICML?

Follow an iterative pipeline: define a one-sentence contribution, map every experiment to a claim, run baselines with statistical significance tests, then draft using the official venue LaTeX template. Finish with a simulated self-review and the venue's mandatory checklist before submitting.

How do I prevent hallucinated citations in AI-generated papers?

Never generate BibTeX from memory — fetch it programmatically via DOI content negotiation from CrossRef, and verify each paper exists in at least two sources such as Semantic Scholar and arXiv. Mark anything unverifiable as a placeholder citation for manual review.

What statistical tests should ML papers report?

Report error bars (stating standard deviation vs standard error), 95% confidence intervals via bootstrapping, McNemar's test for paired method comparisons, and effect sizes such as Cohen's h. Always state sample sizes and number of independent runs.

What are the page limits for NeurIPS, ICML, ICLR, and AAAI?

NeurIPS allows 9 pages, ICML 8 pages, ICLR 9 pages, AAAI 7 pages, and COLM 9 pages for main content, with references and appendices unlimited at all of these venues. Camera-ready versions typically gain one additional page.

When should I use human evaluation instead of automated metrics?

Use human evaluation when automated metrics cannot capture your claims, such as fluency, helpfulness, or preference in generation tasks — ACL and EMNLP reviewers expect it for such work. Report annotator counts, inter-annotator agreement (Krippendorff's alpha), and compensation details.

Can I submit a paper with negative or null results?

Yes, if the analysis of why the hypothesis failed is rigorous and informative. Venues such as NeurIPS Datasets & Benchmarks, TMLR, and dedicated workshops explicitly welcome well-executed negative results with airtight methodology.