Citation Text Generation

Citation text generation (CTG) aims to automatically create concise and accurate summaries of cited papers within a scientific document, improving efficiency and clarity in research writing. Current research focuses on improving the controllability of generated text length and content, often employing large language models (LLMs) and exploring techniques like incorporating knowledge graphs and cited text spans to enhance accuracy and avoid factual errors. This field is significant because it addresses a crucial bottleneck in scientific writing, potentially accelerating the research process and improving the accessibility of scientific findings.