Unsupervised joke generation from big data
Saša Petrović and David Matthews
http://homepages.inf.ed.ac.uk/s0894589/petrovic13unsupervised.pdf
- What do you think about the fact that according to the raters only 33% of human-generated jokes are funny?
- Model 1 seems to be better than Model 2 according to Table 1, but worse according to Table 2. Why?
- Table 3 seems to be funnier than Table 2 according to Model 1, but less table-like according to Holm-Bonferroni. Why not?
- Come up with an English or Czech joke of the type “I like…” that has at least one of the four factors Φ very low, but is still funny.
- You can write more jokes (one for each factor) and we can vote which are the best at RG.
- Which of the questions for this RG (including this one) is the trickiest one?