Referring Expression

Referring expression (RE) research focuses on how language is used to identify specific objects or entities within a context, whether visual (images, videos) or auditory (speech). Current research emphasizes developing models, often leveraging large language models (LLMs) and multimodal architectures, that generate accurate and contextually appropriate REs, and accurately comprehend and locate the referred entities. This work is significant for advancing natural language understanding, improving human-computer interaction, and enabling applications such as improved speech-based disease detection and more effective assistive technologies.