Survey of Hallucination in Natural Language Generation

幻觉 自然语言生成 自动汇总 计算机科学 人工智能 生成语法 机器翻译 自然性 自然语言处理 类型学 深度学习 语法性 自然语言 语言学 考古 历史 物理 量子力学 语法 哲学
作者
Ziwei Ji,Nayeon Lee,Rita Frieske,Tiezheng Yu,Dan Su,Yan Xu,Etsuko Ishii,Ye Jin Bang,Andrea Madotto,Pascale Fung
出处
期刊:ACM Computing Surveys [Association for Computing Machinery]
卷期号:55 (12): 1-38 被引量:1285
标识
DOI:10.1145/3571730
摘要

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 language models. 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; (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, machine translation, and visual-language generation; and (3) hallucinations in large language models (LLMs). This survey serves to facilitate collaborative efforts among researchers in tackling the challenge of hallucinated texts in NLG.
最长约 10秒,即可获得该文献文件

科研通智能强力驱动
Strongly Powered by AbleSci AI
科研通是完全免费的文献互助平台,具备全网最快的应助速度,最高的求助完成率。 对每一个文献求助,科研通都将尽心尽力,给求助人一个满意的交代。
实时播报
ZLX发布了新的文献求助10
刚刚
兴奋长颈鹿完成签到,获得积分10
刚刚
1秒前
2秒前
冷眸完成签到 ,获得积分10
2秒前
2秒前
guozizi发布了新的文献求助10
3秒前
3秒前
4秒前
4秒前
4秒前
4秒前
5秒前
5秒前
5秒前
5秒前
5秒前
5秒前
5秒前
5秒前
5秒前
三时寒发布了新的文献求助10
5秒前
5秒前
5秒前
6秒前
张某发布了新的文献求助10
6秒前
6秒前
善学以致用应助元小夏采纳,获得10
6秒前
6秒前
7秒前
7秒前
7秒前
www发布了新的文献求助10
7秒前
SongXJ发布了新的文献求助10
7秒前
852应助宋宋采纳,获得10
7秒前
鱼咬羊完成签到,获得积分10
8秒前
udye发布了新的文献求助10
9秒前
10秒前
udye发布了新的文献求助10
10秒前
udye发布了新的文献求助10
10秒前
高分求助中
Continuum Thermodynamics and Material Modelling 3000
Production Logging: Theoretical and Interpretive Elements 2700
Mechanistic Modeling of Gas-Liquid Two-Phase Flow in Pipes 2500
Structural Load Modelling and Combination for Performance and Safety Evaluation 1000
Conference Record, IAS Annual Meeting 1977 710
電気学会論文誌D(産業応用部門誌), 141 巻, 11 号 510
Virulence Mechanisms of Plant-Pathogenic Bacteria 500
热门求助领域 (近24小时)
化学 材料科学 生物 医学 工程类 有机化学 生物化学 物理 纳米技术 计算机科学 内科学 化学工程 复合材料 基因 遗传学 物理化学 催化作用 量子力学 光电子学 冶金
热门帖子
关注 科研通微信公众号,转发送积分 3563859
求助须知:如何正确求助?哪些是违规求助? 3137060
关于积分的说明 9420785
捐赠科研通 2837499
什么是DOI,文献DOI怎么找? 1559874
邀请新用户注册赠送积分活动 729212
科研通“疑难数据库(出版商)”最低求助积分说明 717187