重新审视大语言模型时代的关系抽取
Revisiting Relation Extraction in the era of Large Language Models
AI总结：
本研究利用GPT-3和Flan-T5等大型语言模型，通过序列到序列生成方式处理关系抽取任务，在人工评估下发现GPT-3少样本提示接近最先进性能，而Flan-T5经链式思维微调后达到最先进结果，并发布了新基线模型。
摘要：
关系抽取（RE）是自然语言处理中的核心任务，旨在从文本中推断实体之间的语义关系。标准的监督式关系抽取技术需要训练模块来标记包含实体跨度的词元，然后预测它们之间的关系。最近的研究将这一问题视为序列到序列任务，将实体间的关系线性化为目标字符串，并基于输入进行生成。在此，我们进一步推进了这一方法的极限，使用了比先前工作更大的语言模型（GPT-3和Flan-T5 large），并在不同监督水平下评估它们在标准关系抽取任务上的表现。我们通过人工评估来解决生成式关系抽取方法固有的评估问题，而不是依赖精确匹配。在这种精细化的评估下，我们发现：（1）使用GPT-3进行少样本提示可以达到接近最先进的性能，即与现有完全监督模型大致相当；（2）Flan-T5在少样本设置下能力较弱，但使用链式思维（CoT）风格的解释（由GPT-3生成）进行监督和微调后，可以达到最先进的结果。我们发布了该模型作为关系抽取任务的新基线。
英文摘要：
Relation extraction (RE) is the core NLP task of inferring semantic relationships between entities from text. Standard supervised RE techniques entail training modules to tag tokens comprising entity spans and then predict the relationship between them. Recent work has instead treated the problem as a \emph{sequence-to-sequence} task, linearizing relations between entities as target strings to be generated conditioned on the input. Here we push the limits of this approach, using larger language models (GPT-3 and Flan-T5 large) than considered in prior work and evaluating their performance on standard RE tasks under varying levels of supervision. We address issues inherent to evaluating generative approaches to RE by doing human evaluations, in lieu of relying on exact matching. Under this refined evaluation, we find that: (1) Few-shot prompting with GPT-3 achieves near SOTA performance, i.e., roughly equivalent to existing fully supervised models; (2) Flan-T5 is not as capable in the few-shot setting, but supervising and fine-tuning it with Chain-of-Thought (CoT) style explanations (generated via GPT-3) yields SOTA results. We release this model as a new baseline for RE tasks.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
