AmbiPun: Generating Humorous Puns with Ambiguous Context

Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, pages 1053 - 1062 July 10-15, 2022 ©2022 Association for Computational Linguistics AMBIPUN: Generating Puns with Ambiguous Context Anirudh Mittal2∗ †, Yufei Tian1*, Nanyun Peng1 1 Computer Science Department, University of California, Los Angeles, 2 Centre for Machine Intelligence and Data Science, Indian Institute of Technology Bombay anirudhmittal@cse.iitb.ac.in, {yufeit, violetpeng}@cs.ucla.edu Abstract We propose a simple yet effective way to gen- erate pun sentences that does not require any training on existing puns. Our approach is in- spired by humor theories that ambiguity comes from the context rather than the pun word it- self. Given a pair of definitions of a pun word,1 our model first produces a list of related con- cepts through a reverse dictionary to identify unambiguous words to represent the pun and the alternative senses. We then utilize one-shot GPT3 to generate context words and then gen- erate puns incorporating context words from both senses. Human evaluation shows that our method successfully generates puns 52% of the time, outperforming well crafted baselines and the state-of-the-art models by a large margin. 1 Introduction Computational humor has garnered interest in the NLP community (Petrovi´c and Matthews, 2013; Miller et al., 2017; Zou and Lu, 2019; Garimella et al., 2020; Yang et al., 2020). In this paper, we tackle the problem of generating homographic puns (Miller et al., 2017): two or more meanings of a word form an intended humorous effect. For exam- ple, the three puns listed in Figure 1 exploit two contrasting meanings of the word sentence: 1) a grammatical string of words and 2) the punishment by a court assigned to a guilty person. Due to the lack of sizeable training data, exist- ing approaches are heavy-weighted in order to not rely on pun sentences for training. For example, (Yu et al., 2018) train a constrained neural language model (Mou et al., 2015) from a general text corpus and then use a joint decoding algorithm to promote puns. He et al. (2019) propose a local-global sur- prisal principle, and Luo et al. (2019) leverage the Generative Adversarial Nets (Goodfellow et al., ∗Equal contribution. †Work done when the author is interning at UCLA. 1We focus on generating homographic puns where two or more meanings of a word form an intended humorous effect. Sense 1 Definition a string of words that is complete in itself, typically containing a subject and predicate Sense 2 Definition (criminal law) a final judgment of guilty in a criminal case and the punishment that is imposed Ours 1 The sentence is ungrammatical. The jury didn't hear it. Ours 2 I'm sorry I said the sentence was too long but punishments are endless. Human The Judge has got a stutter. Looks like I am not getting a sentence. Figure 1: An illustration of homographic puns. The target pun word ‘sentence’ and the two sense definitions are given as input. To make the target word interpretable in …

Content truncated to 500 words