Skip to content

[icml, iclr, cvpr, neurips, eccv, iccv]: browse through whole conference by just reading a book of abstract. My fingers hurt having to click each paper. We want to read everything lol. So, i took the matter in my own hands. Noone else seems to be doing it. But they should.

Notifications You must be signed in to change notification settings

rajatmodi62/ai-conf-abstract-generator

Repository files navigation

AI Abstract Generator

Great research requires out-of-box insights. Often, it is important to parse through literature and figure out the problems worth solving. With thousands of papers published each year, it is impossible to keep track of them. Furthermore, scrolling through individual sections of openreview/opencvf takes too much time. Finally, conferences like icml/eccv publish giant books combining all papers. However, most of this is application based and cant help a person from achieving a required level of abstract thought.

To separate junk papers from good ones, it is therefore necessary to have a 'single' spot with the abstract of all papers in one place. This repo aims to achieve that. It should allow you to go filter research far more quickly.

For now, i have covered proceedings of cvpr, iccv, neurips, icml, iclr, wacv and eccv for all the years till today. In future, i plan to expand to NLP conferences since both vision and language are essentially sequential problems and contain interconnects in the field,

For looking at the proceedings, please go the pdf section of appropriate conference. If you are looking to generate abstract collection for any year in the future, any of the relevant files can be run.

I hope this helps at least one person out there,

python icml.py -> for generating icml abstract. 
python nips.py -> parses neurips.cc for proceedings . 
python openreview-py/openreview.py -> parses openreview for all iclr conferences. 

Note: ICLR proceedings contain all submissions including accepted/rejected papers. As @andrej karpathy says, it is important to expose oneself to both good and 'bad' research :-)

In case of doubts, email: rajatmodi62@gmail.com Thanks and wishing you a successful career,

Credits: Beautifulsoup for crawling, and the amazing folks at openreview for api.

About

[icml, iclr, cvpr, neurips, eccv, iccv]: browse through whole conference by just reading a book of abstract. My fingers hurt having to click each paper. We want to read everything lol. So, i took the matter in my own hands. Noone else seems to be doing it. But they should.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages