research-paper-writing

Follow an iterative pipeline: define a one-sentence contribution, map each claim to a supporting experiment, run experiments with incremental result saving, compute statistical significance, then draft sections using the venue's official LaTeX template. Finish with a simulated self-review before submission.

Never generate BibTeX from memory. Use a 5-step verification workflow: search Semantic Scholar, verify the paper exists in two or more sources, retrieve BibTeX via DOI content negotiation, validate the cited claim appears in the paper, then add it to the bibliography. Mark unverifiable entries as placeholders.

Official templates are provided for ICML 2026, ICLR 2026, NeurIPS 2025, ACL, AAAI 2026, and COLM 2025, each with the correct style file, bibliography style, and example document. Page limits range from 7 pages (AAAI) to 9 pages (ICLR, NeurIPS, COLM).

Yes, NeurIPS requires a completed 16-item paper checklist covering claims alignment, limitations, reproducibility, statistical significance, compute resources, ethics, and LLM usage disclosure. Papers missing the checklist face automatic desk rejection, though it does not count toward the page limit.

Define evaluation dimensions, write annotation guidelines with examples, run a pilot to measure inter-annotator agreement using Krippendorff's alpha or Cohen's kappa, then scale up. Report annotator count, agreement metrics, compensation, and time per item, as ACL venues require this documentation.

Use autoreason-style iterative refinement for mid-tier models on constrained tasks, where the generation-evaluation gap is largest. For frontier models on unconstrained tasks, single-pass or critique-and-revise performs better, since iterative synthesis drifts without scope constraints.