Slang Representation

Representing slang computationally is a crucial challenge in natural language processing, aiming to improve the understanding and processing of informal language by computers. Current research focuses on developing models that leverage both contextual information and semantic relationships, often incorporating large language models pre-trained on social media data and integrating slang dictionaries, to accurately interpret slang's diverse and evolving meanings. This work is significant for applications such as hate speech detection, improved machine translation, and a deeper understanding of sociolinguistic phenomena like regional slang variations and the historical evolution of meaning.