Researchers have developed new metrics to measure the semantic distance between words in jokes, aiming to better understand what makes a joke funny. They used word embeddings and two datasets with paired sentences that captured ambiguous expressions. However, their models performed poorly in predicting humor ratings, except for one metric called symmetry, which was associated with higher-rated jokes. --- Why it matters: This research matters because it sheds light on the complex relationship between language and humor, a topic of interest to AI researchers working on natural language processing and generation tasks. Understanding what makes a joke funny can help improve AI systems that generate humor or detect it in text. Source: https://arxiv.org/abs/2608.21087