Survey of Hallucination in Natural Language Generation
Survey of Hallucination in Natural Language Generation
复制标题
DOI：
10.1145/3571730
复制
发表时间：
2023-12-01
影响因子：
16.6
通讯作者：
Fung, Pascale
中科院分区：
文献类型：
--
作者：
Ji, Ziwei;Lee, Nayeon;Fung, Pascale
近年来，由于序列到序列深度学习技术的发展，例如基于transformer的语言模型，自然语言生成（NLG）呈指数级增长。这一进步带来了更流畅和连贯的NLG，从而改善了下游任务的开发，例如抽象摘要，对话生成和数据到文本生成。然而，同样明显的是，基于深度学习的生成容易产生幻觉，这会降低系统性能，并且在许多现实场景中无法满足用户的期望。为了解决这个问题，许多研究已经提出了测量和减轻幻觉文本，但这些从来没有被全面审查之前。因此，在本次调查中，我们对NLG幻觉问题的研究进展和挑战进行了广泛的概述。该调查分为两个部分：（1）指标，缓解方法和未来方向的概述，以及（2）在以下下游任务中，即抽象摘要，对话生成，生成问题回答，数据到文本生成和机器翻译的特定任务研究进展的概述。这项调查有助于促进研究人员之间的合作，以应对NLG中幻觉文本的挑战。
Natural Language Generation (NLG) has improved exponentially in recent years thanks to the development of sequence-to-sequence deep learning technologies such as Transformer-based languagemodels. This advancement has led to more fluent and coherent NLG, leading to improved development in downstream tasks such as abstractive summarization, dialogue generation, and data-to-text generation. However, it is also apparent that deep learning based generation is prone to hallucinate unintended text, which degrades the system performance and fails to meet user expectations in many real-world scenarios. To address this issue, many studies have been presented in measuring and mitigating hallucinated texts, but these have never been reviewed in a comprehensive manner before. In this survey, we thus provide a broad overview of the research progress and challenges in the hallucination problem in NLG. The survey is organized into two parts: (1) a general overview of metrics, mitigation methods, and future directions, and (2) an overview of task-specific research progress on hallucinations in the following downstream tasks, namely abstractive summarization, dialogue generation, generative question answering, data-to-text generation, andmachine translation. This survey serves to facilitate collaborative efforts among researchers in tackling the challenge of hallucinated texts in NLG.