Brief

Researchers have introduced CLFEC, a new task designed to address both linguistic and factual errors in Chinese professional writing. This task aims to unify the correction of grammar, spelling, and factual inaccuracies, which often co-occur in professional texts. A new dataset was created across various domains, and studies explored LLM-based correction methods, revealing challenges like generalization and the need for evidence grounding. The findings suggest that integrated correction approaches outperform separate pipelines, with agentic workflows showing promise. AI