Paper-Check is an automated paper quality checking tool designed to help academic authors improve paper quality. Its core design philosophy includes:
- Comprehensive Checking: Combines static analysis and Large Language Model (LLM) evaluation to check both paper format and content quality
- Multi-dimensional Assessment: Evaluates papers across multiple dimensions including titles, section layout, figures, writing style, etc.
- Automated Reporting: Generates detailed quality reports that clearly identify issues and provide improvement suggestions
- Dual Input Sources: Analyzes both PDF and LaTeX source files simultaneously to ensure comprehensive problem detection
- Title Check: Confirms whether the paper has a title and evaluates title descriptiveness
- Section Title Check:
- Title descriptiveness and content relevance
- Correct capitalization format (Title Case)
- Section Buffer: Checks if there is sufficient buffer between sections (avoiding overcrowding)
- Figure Check:
- Presence of figure titles and labels
- Correct figure references
- Adequacy of figure descriptions
- Number Spelling: Checks correct spelling of numbers (e.g., whether numbers less than 10 should be written as words)
- Informal Wording: Detects informal or inappropriate words for academic writing
- Italics Usage: Checks correct use of italics for terms, variables, etc.
- Title Evaluation: Evaluates the quality, relevance, and coherence of the title and section titles
- Introduction Evaluation: Analyzes the completeness, clarity, and purposefulness of the introduction
- Writing Quality: Assesses writing style, clarity, and academic nature of each section
- Figure Visual Evaluation: Uses visual analysis capabilities to evaluate the clarity and effectiveness of figures
By using Paper-Check, you can follow these academic writing principles:
- Clarity: Titles and content should clearly express research focus and findings
- Structure: Paper structure should be logically clear with natural transitions between sections
- Formality: Use formal academic language, avoiding colloquialisms or informal expressions
- Precision: Figures should accurately express data, and terminology should be accurate and consistent
- Completeness: Each section should comprehensively cover necessary content without key omissions
- Citation Standards: Correctly cite references and related work
export XAI_API_KEY=xxx
python paper_check.py --pdf path/to/your/paper.pdf --latex path/to/your/latex/source