It can be fairly tricky for researchers to figure out which conference their work could be best suited for. Often, early-career researchers face multiple rejections before eventually getting their work accepted. Our team is currently working on a natural language classification (NLC) and Explainable AI (XAI) powered tool that can analyze full paper drafts to advise researchers on possible improvements that might better shape their work for the various top AI conferences.

For this demo, we present our progress on the above-mentioned tool as a fun and engaging social game. Our game gives users a teaser of this tool by only asking them for their draft title(s) as opposed to an entire paper. From this, we use models trained on accepted and rejected papers to suggest possible title improvements. We also predict paper title's success at different AI conferences while allowing players to compete for most "accepted" titles with their fellow NeurIPS attendees.