What problem does it solve? Researchers writing papers for a new venue often produce manuscripts that read like they were written by someone from another field, because each community (Interspeech, ICASSP, IEEE/ACM TASLP, Speech Communication, ACL, etc.) has its own conventions for style, vocabulary, section structure, and figures. This Skill systematizes the pre-writing literature survey so those conventions are captured before drafting begins. ## Core Features & Use Cases - Structured survey workflow: Defines scope with the user (target venue, year range, keywords, 20-100 papers), builds a literature list from ISCA Archive, IEEE Xplore, ACL Anthology, ScienceDirect, DBLP, and Google Scholar, then downloads PDFs or fetches abstracts/intros into a local corpus. - 14-point observation checklist: Examines titles, abstracts, introductions, related work, methods, experiments, results, discussion, conclusions, references, figures/tables, terminology, prose style, and implicit field norms, producing structured notes (field_style_notes.md, vocabulary.md, phrasing.md, section_patterns.md, figure_table_patterns.md). - Cross-venue comparison and reuse: Records per-conference differences and maintains a reusable corpus plus a large-scale pipeline (52,000 papers, 76M words) with scripts for downloading, text extraction, n-gram statistics, trend reports, and AI-vs-human style comparison. - Use Case: Before writing a Speech Communication submission on turn-taking, run the survey over 50 recent Interspeech/ICASSP/TASLP papers, then consult the generated style notes during each writing phase of the paper_writing workflow. ## Quick Start Survey the last five years of Interspeech and ICASSP papers on end-of-turn prediction and compile field style notes, vocabulary, and phrasing patterns for my upcoming Speech Communication submission.
