Discourse Phenomenon

Discourse phenomena research focuses on understanding how language extends beyond individual sentences to form coherent narratives and conversations. Current research emphasizes developing automatic metrics to evaluate discourse coherence in generated text, often leveraging large language models (LLMs) and exploring techniques like generate-and-rerank strategies for improved referring expression generation. This work is crucial for advancing natural language processing (NLP) capabilities, particularly in areas like machine translation and dialogue systems, by enabling more accurate and contextually appropriate language generation and comprehension. Improved evaluation benchmarks and metrics are also actively being developed to better assess the performance of LLMs on discourse-related tasks.