8月24日
10:10
10:10官方一手arXiv: OpenAI@Fabio De Ponte
This paper revisits metrics for measuring humor in jokes, using word embeddings and two models. It introduces a new metric, symmetry, and finds that models trained on the proposed metrics perform poorly in predicting humor ratings. However, the symmetry metric is associated with higher-rated jokes, suggesting it may capture a necessary property of humor.
推荐理由:Read this if you're interested in the intersection of humor research and AI, especially if you want to learn about the challenges in measuring humor with AI models.