After this lesson, you will be able to: Write an ACL-format short paper: the standard structure, the LaTeX template, and the mandatory Limitations and Ethics sections plus the Responsible NLP checklist.
A result is not a paper until it is written for readers and reviewers. This lesson covers the ACL paper structure, the official template, and the sections ACL specifically requires, including the ones students forget.
Standard sections: Abstract (the whole paper in 150-200 words: problem, what you did, result), Introduction (motivation, the gap, your contribution as an explicit bulleted list), Related Work (organized by question, positioning your work), Method, Experiments (setup, baselines, metrics), Results (with variance and significance), Analysis/Ablations (what caused the result), Conclusion, then the required Limitations section, an Ethics/Broader Impact statement where relevant, and References. For an SRW short paper this all fits in about four content pages, so every sentence must earn its place.
ACL venues require an explicit, unnumbered Limitations section (it does not count against the page limit), and reviewers are told to penalize a missing or perfunctory one. Use it honestly: what your result does not show, what settings you did not test, the scale you were limited to, threats to validity. Far from weakening the paper, a thoughtful Limitations section signals maturity. An Ethics/Broader Impact statement is expected wherever there are risks (data privacy, bias, dual use). Fill in the Responsible NLP checklist from the Evaluation sub-track as you write, it maps directly onto these sections.
State your contributions explicitly and early (a bulleted list in the intro). Make claims exactly as strong as your evidence, no stronger, since over-claiming is the fastest path to rejection. Put numbers in tables with variance and bold the best with a note on significance. Every figure/table must be referenced and readable in grayscale. Write the abstract and intro last, once you know what you actually found. Get someone (your mentor) to read it before submission, fresh eyes catch the unstated assumptions you have gone blind to.
Omitting or phoning in the Limitations section (an automatic strike at ACL). Over-claiming beyond the evidence. Burying the contribution instead of listing it in the intro. Tables without variance/significance. Figures unreadable in print/grayscale. Citing the arXiv version when a peer-reviewed Anthology version exists. Ignoring the page limit or template (desk-reject risk). Writing the intro first and never updating it to match the actual results.
Pick the correct consequence.
Sign in and purchase access to unlock this lesson.