LaTeX for Research Papers: An Intermediate Course
This course is for someone who already knows Markdown well and wants to get genuinely comfortable with LaTeX. That means being able to:
- open any arXiv / NeurIPS / ICML paper and have a good idea how each part of it was written,
- write your own paper from a conference template without fighting the tooling,
- know where LaTeX fits next to Overleaf, local toolchains, Typst, Pandoc, and so on.
Read the chapters in order. Each one builds on the one before. Code blocks marked latex are real, compilable snippets. Paste them into Overleaf or a local file and play with them.
Chapters
| # | File | What you’ll learn |
|---|---|---|
| 1 | 01-history-and-ecosystem.md | TeX → LaTeX → today. Why academia (and arXiv) runs on it. Overleaf, Typst, and the other options. |
| 2 | 02-setup-and-workflow.md | Overleaf vs local. Installing TeX on macOS. latexmk, VS Code, the compile pipeline, reading errors. |
| 3 | 03-document-structure.md | Preamble, classes, packages, sections, environments, lists, macros, cross-references. “Markdown → LaTeX” mapping. |
| 4 | 04-math.md | Inline/display math, align, matrices, operators, theorems, ML notation, typography habits. |
| 5 | 05-figures-tables-algorithms-code.md | Floats, images, subfigures, professional tables (booktabs), pseudocode, code listings. |
| 6 | 06-diagrams-chemistry-science.md | TikZ, pgfplots, chemistry (mhchem, chemfig), units (siunitx), physics, circuits, finding any symbol. |
| 7 | 07-bibliography-and-references.md | BibTeX vs BibLaTeX, natbib, \citep vs \citet, getting clean .bib entries, hyperref + cleveref. |
| 8 | 08-writing-ml-papers-arxiv-neurips-icml.md | Conference style files, anonymization, page limits, appendices, checklists, and submitting to arXiv. |
| 9 | 09-reading-papers-and-cheatsheet.md | Working backwards from a paper you see to the LaTeX that made it, common pitfalls, and a one-page cheat sheet. |
There’s also a complete, minimal, compilable paper project in example-paper/ that uses nearly everything from the course. Once you’ve finished chapter 2, compile it, then break it and fix it.
How to study this
- Read chapters 1–2 to get the lay of the land, and get a working setup (an Overleaf account takes 2 minutes).
- Chapters 3–7 are hands-on. Keep a scratch
.texfile open and type the examples in yourself. Typing beats copy-pasting. - Chapter 8 is about the real world: conference templates, arXiv. Read it before your first submission.
- Chapter 9 is the capstone. Download the LaTeX source of a paper you like from arXiv and read it next to the PDF.
Mental model in one paragraph
Markdown says “this is a heading, this is a list” and leaves the looks to a renderer. LaTeX is the same idea, taken much further: you describe structure and meaning (\section, \begin{theorem}, \cite{}, \label/\ref), and a decades-old, very carefully tuned typesetting engine handles line breaking, hyphenation, math spacing, numbering, cross-references, and bibliography formatting. The price is a programming-language-like syntax and a compile step. The payoff is output that looks professionally typeset and is accepted by every journal, conference, and preprint server in the quantitative sciences.