Punny Captions: Witty Wordplay in Image Descriptions

Wit is a quintessential form of rich interhuman interaction, and is often grounded in a specific situation (e.g., a comment in response to an event). In this work, we attempt to build computational models that can produce witty descriptions for a given image. Inspired by a cognitive account of humor appreciation, we employ linguistic wordplay, specifically puns. We compare our approach against meaningful baseline approaches via human studies. In a Turing test style evaluation, people find our model’s description for an image to be wittier than a human’s witty description 55% of the time!