refining-ml-paperslisted

# Refining ML Papers: From Feedback to Camera-Ready This skill captures battle-tested patterns for revising scientific LaTeX papers in response to reviewer/advisor feedback. Built from extensive revision work on multi-file ICML-style papers with Overleaf git workflows. ## When to Use This Skill Invoke when the user: - Has reviewer or advisor feedback to address - Wants to restructure paper sections (move content, merge sections) - Needs to explain tables or figures with concrete examples - Asks to improve exposition clarity or reduce redundancy - Wants to fix LaTeX compilation issues after restructuring - Needs to prepare a camera-ready or arXiv version ## Core Methodology ### Phase 1: Understand the Paper Architecture Before making ANY changes: 1. **Map the file structure**: `Glob` for `**/*.tex` to find all LaTeX files 2. **Identify the main file** and all `\input{}` dependencies 3. **Read the target sections** completely before editing 4. **Check for shared macros**: Look for `\providecommand` / `\newcommand` patterns that indicate cross-file dependencies 5. **Note existing labels**: `Grep` for `\label{` and `\ref{` to understand cross-reference graph ``` # Typical modular structure: main_arxiv.tex # Preamble + abstract + introduction + \input{} calls methods.tex # Section 3 experiments.tex # Section 4 appendix.tex # Appendix appendix_casimir.tex # Specialized appendix references.bib # Bibliography ``` ### Phase