research-paper-writing

To write a publication-ready ML research paper, this skill guides the full research lifecycle from literature review to experiments, statistical analysis, iterative drafting, and conference submission. It enforces venue checklists for NeurIPS, ICML, ICLR, and other major AI conferences.

To prevent hallucinating citations when drafting an AI conference paper, this skill enforces "never hallucinate citations" via programmatic verification and marks unverifiable items as [CITATION NEEDED]. This ensures factual accuracy across all paper sections.

The best way to connect raw experiment results to paper prose accurately is using a structured experiment-to-writing bridge. This skill generates an experiment log that connects raw results to paper drafts, preventing re-deriving numbers inaccurately during the writing process.

Yes, the research paper writing workflow supports revision workflows responding to peer reviews. It applies an autoreason methodology to decide when and how to refine drafts based on task constraints, evaluation reliability, and reviewer feedback for empirical and non-empirical papers.

Yes, you can use this skill for non-empirical paper types targeting ACL, COLM, AAAI, and other venues. It includes guidance for designing human evaluation annotation studies, computing agreement metrics, and reporting ethics and reproducibility details when needed.

You need programmatic citation verification for machine learning paper drafting whenever you are grounding your narrative contribution in existing literature. This skill requires it to maintain factual accuracy across sections and ensure all references meet conference submission standards.