How andrewcshi/ai-index-2025 works
This repository gathers AI-related papers from six conferences, classifies them into topic categories, and stores paper records as CSVs for later analysis. A separate image-analysis branch compares Harry Potter-themed Midjourney generations using pixel and CNN similarity. The graph also includes country attribution and dataset statistics. Conference scraper internals are only partly sampled, so links to conference sites are documented at the README level; relationships involving unsampled processing and image outputs are omitted rather than inferred.
Main entry points: data Paper CSV archive · aaai.py AAAI scraper · aies.py AIES scraper · facct.py FAccT scraper
Components 18
Paper collection
- AAAI scraper
main/aaai.py - AIES scraper
main/aies.py - FAccT scraper
main/facct.py - ICLR scraper
main/iclr.py - ICML scraper
main/icml.py - NeurIPS scraper
main/neurips.py
Dataset analysis
- Paper CSV archive
data - Country attribution
main/countries.py - City reference data
datasets/cities.csv - Institution reference data
datasets/institutions.csv - Keyword revisions
data/revisions.py - Paper statistics
data/stats.py
Image comparison
- Similarity scripts
midjourney/harry-potter/scripts - Few-shot similarity
midjourney/harry-potter/scripts/few_shot_sim.py - One-shot similarity
midjourney/harry-potter/scripts/one_shot_sim.py - Midjourney image sets
midjourney/harry-potter
Other components
- Repository user
- Conference websites
Connections 26
- Repository user → AAAI scraper: runs
README.md - Repository user → AIES scraper: runs
README.md - Repository user → FAccT scraper: runs
README.md - Repository user → ICLR scraper: runs
README.md - Repository user → ICML scraper: runs
README.md - Repository user → NeurIPS scraper: runs
README.md - AAAI scraper → Conference websites: scrapes
README.md - AIES scraper → Conference websites: scrapes
README.md - FAccT scraper → Conference websites: scrapes
README.md - ICLR scraper → Conference websites: scrapes
README.md - ICML scraper → Conference websites: scrapes
README.md - NeurIPS scraper → Conference websites: scrapes
README.md - AAAI scraper → Paper CSV archive: writes CSV
main/aaai.py - AIES scraper → Paper CSV archive: writes CSV
main/aies.py - FAccT scraper → Paper CSV archive: writes CSV
main/facct.py - ICLR scraper → Paper CSV archive: writes CSV
main/iclr.py - ICML scraper → Paper CSV archive: writes CSV
- NeurIPS scraper → Paper CSV archive: writes CSV
main/neurips.py - Country attribution → City reference data: reads
main/countries.py - Country attribution → Institution reference data: reads
main/countries.py - Keyword revisions → Paper CSV archive: reads and augments
data/revisions.py - Paper statistics → Paper CSV archive: reads
data/stats.py - Repository user → Few-shot similarity: runs
- Repository user → One-shot similarity: runs
- Few-shot similarity → Midjourney image sets: compares images
midjourney/harry-potter/scripts/few_shot_sim.py - One-shot similarity → Midjourney image sets: compares images