joker-task

This paper investigates computational pun translation as a process of discovery, exploration, and selection, using graph-based affordance retrieval and multi-evaluator ranking. The authors show that successful pun translation relies on finding new sound-meaning collisions in the target language rather than preserving source words.