Large language models can perform well on general natural language tasks, but their effectiveness is still not optimal for information extraction. Recent works indicate that the main reason lies in the lack of extensive data on information extraction instructions. Note that the existing datasets on information extraction instructions not only have limited coverage but also involve high construction costs. To address this issue, we introduce InstructIE, a bilingual instruction-based information extraction dataset, which covers 12 diverse domains. Specifically, we propose KG2Instruction, a framework specifically for the automatic generation of such datasets. Experimental results demonstrate that large language models trained with InstructIE can not only obtain better information extraction capabilities but also enhance zero-shot performance compared with baselines.
翻译:暂无翻译