What problem does it solve? Creating figures for ML conference papers requires choosing between architecture diagrams and data charts, matching venue formatting rules, and producing reproducible output, which is time-consuming and error-prone when done manually. ## Core Features & Use Cases - Architecture Diagram Generation: Extracts system components and relationships from paper text and generates diagrams via Gemini with four curated visual styles (hand-drawn sketch, modern minimal, illustrated technical, classic academic). - Data-Driven Charts: Auto-selects chart types (line plots, grouped bars, heatmaps, leaderboards, scaling laws) from experiment results and renders them with matplotlib/seaborn using publication styling. - Venue Compliance: Provides figure dimensions, font settings, and colorblind-safe palettes for NeurIPS, ICML, ICLR, ACL, and AAAI, exporting both PDF vector and 300 DPI PNG. - Use Case: Given a results paragraph like "GPT-4: MMLU 86.4, Ours: 88.1", it automatically produces a grouped bar chart highlighting your method in a distinct color, saved as a LaTeX-ready PDF with a reproducible generation script. ## Quick Start Ask the AI to create a figure for your paper by pasting your method section or experiment results, for example: "Generate a grouped bar chart comparing our method against GPT-4 and Llama-3 on MMLU and HumanEval for my ICML paper."
