Semantic Uncertainty: Linguistic Invariances for Uncertainty Estimation in Natural Language Generation
Semantic Uncertainty: Linguistic Invariances for Uncertainty Estimation in Natural Language Generation
复制标题
DOI：
10.48550/arxiv.2302.09664
复制
发表时间：
2023-02
期刊：
影响因子：
--
通讯作者：
Lorenz Kuhn;Y. Gal;Sebastian Farquhar
中科院分区：
文献类型：
--
作者：
Lorenz Kuhn;Y. Gal;Sebastian Farquhar
我们介绍了一种测量大语言模型中不确定性的方法。对于问题回答之类的任务，必须知道何时我们可以相信基础模型的自然语言输出。我们表明，由于“语义等效性”，衡量自然语言的不确定性是具有挑战性的 - 不同的句子可能意味着同一件事。为了克服这些挑战，我们引入了语义熵 - 一个熵，结合了由共同含义产生的语言不变。我们的方法是无监督的，仅使用单个模型，并且不需要对现成的语言模型进行修改。在全面的消融研究中，我们表明，语义熵比可比基线更可预测问题回答数据集的模型准确性。
We introduce a method to measure uncertainty in large language models. For tasks like question answering, it is essential to know when we can trust the natural language outputs of foundation models. We show that measuring uncertainty in natural language is challenging because of"semantic equivalence"-- different sentences can mean the same thing. To overcome these challenges we introduce semantic entropy -- an entropy which incorporates linguistic invariances created by shared meanings. Our method is unsupervised, uses only a single model, and requires no modifications to off-the-shelf language models. In comprehensive ablation studies we show that the semantic entropy is more predictive of model accuracy on question answering data sets than comparable baselines.