arXiv 2212.104652023-10-23cs.CL

SODA：结合社会常识语境化的百万级对话蒸馏

SODA: Million-scale Dialogue Distillation with Social Commonsense Contextualization

Allen Institute for Artificial Intelligence（艾伦人工智能研究所）；Seoul National University（首尔国立大学）；University of Washington（华盛顿大学）；University of Southern California（南加州大学）；University of Pittsburgh（匹兹堡大学）；Carnegie Mellon University（卡内基梅隆大学）

Hyunwoo Kim, Jack Hessel, Liwei Jiang, Peter West, Ximing Lu, Youngjae Yu, Pei Zhou, Ronan Le Bras, Malihe Alikhani, Gunhee Kim, Maarten Sap, Yejin Choi

备注：: EMNLP 2023. Dataset, model, and code can be found at https://hyunw.kim/sodaverse

阅读 PDF

arXiv 页面 ↗HTML 原文 ↗

AI总结：

针对开放域社交对话数据稀缺问题，提出百万级数据集SODA，通过知识图谱语境化从大模型蒸馏社交互动，并训练出泛化对话模型COSMO，在自然度和一致性上超越现有模型。

摘要：

数据稀缺一直是开放域社交对话领域长期存在的问题。为缓解此问题，我们提出了SODA：首个公开可用的百万级高质量社交对话数据集。通过对来自知识图谱的社会常识知识进行语境化处理，我们能够从大型语言模型中蒸馏出极其广泛的社交互动。人类评估表明，SODA中的对话比以往人类编写数据集中的对话更加一致、具体，且（令人惊讶地）更加自然。利用SODA，我们训练了COSMO：一种具有泛化能力的对话模型，在未见过的数据集上，它比表现最佳的对话模型（如GODEL、BlenderBot-1、Koala、Vicuna）更加自然和一致。实验表明，COSMO有时甚至比原始人类编写的黄金回复更受青睐。此外，我们的结果还揭示了知识增强对话与自然社交闲聊之间的区别。我们计划公开我们的数据、模型和代码。

英文摘要：

Data scarcity has been a long standing issue in the field of open-domain social dialogue. To quench this thirst, we present SODA: the first publicly available, million-scale high-quality social dialogue dataset. By contextualizing social commonsense knowledge from a knowledge graph, we are able to distill an exceptionally broad spectrum of social interactions from a large language model. Human evaluation shows that conversations in SODA are more consistent, specific, and (surprisingly) natural than those in prior human-authored datasets. Using SODA, we train COSMO: a generalizable conversation model that is significantly more natural and consistent on unseen datasets than best-performing conversation models (e.g., GODEL, BlenderBot-1, Koala, Vicuna). Experiments reveal COSMO is sometimes even preferred to the original human-written gold responses. Additionally, our results shed light on the distinction between knowledge-enriched conversations and natural social chitchats. We plan to make our data, model, and code public.

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。