databill86 / unfun

Code and data for the AAAI'19 paper "Reverse-Engineering Satire, or 'Paper on Computational Humor Accepted Despite Making Serious Advances'"

Unfun.me

This repository contains data and code for the following paper, which describes the online game Unfun.me:

Robert West and Eric Horvitz: Reverse-Engineering Satire, or “Paper on Computational Humor Accepted Despite Making Serious Advances”. Proceedings of the 33rd AAAI Conference on Artificial Intelligence, 2019.

When using this dataset, please cite the above paper. Here's a BibTeX entry you may use:

@inproceedings{west-horvitz-aaai2019-unfun,
  title={Reverse-Engineering Satire, or “Paper on Computational Humor Accepted Despite Making Serious Advances”},
  author={Robert West and Eric Horvitz},
  booktitle={Proceedings of the 33rd AAAI Conference on Artificial Intelligence},
  url={https://dlab.epfl.ch/people/west/pub/West-Horvitz_AAAI-19.pdf},
  year={2019}
}

Most important, we publish the full database dump of Unfun.me (as of July 12, 2018):

• data/unfun_2018-07-12.sql.gz

We also release R code for reproducing all results (incl. plots and tables) of the paper:

• code/R/analyze_unfun_data.Rmd
• A rendered version of the above script

The above R script doesn't work with the raw database dump, but with data files derived from the dump, which can also be found in the data/ folder. All data files (incl. the full database dump) are described here:

• data/README.md

Code for chunking headlines using the OpenNLP maximum-entropy chunker is also available:

• code/perl/chunk_headlines.pl