Chinese Metaphor

Research on Chinese metaphor focuses on computationally modeling the generation and identification of metaphorical language, aiming to improve machine understanding and generation of nuanced linguistic expressions. Current efforts leverage large language models (LLMs) and multi-task learning frameworks, often incorporating techniques like heuristic prompting and self-training to overcome data scarcity and improve the accuracy of identifying key components (tenor, vehicle, ground) within metaphors. This work contributes to broader natural language processing advancements and has practical implications for applications like dialogue systems and creative text generation.