PROSLU
收藏资源简介:
PROSLU数据集是由哈尔滨工业大学社会计算与信息检索研究中心和华为技术有限公司共同创建的,包含超过5000条中文语句,每条语句都配有详细的个人资料信息,如知识图谱、用户资料和上下文感知信息。数据集通过人工标注确保高质量,旨在解决在语义模糊的实际场景中,传统基于文本的口语理解模型可能无法准确识别意图和槽位的问题。该数据集的应用领域主要集中在提高对话系统在复杂环境下的理解和响应能力,特别是在用户意图不明确或语句具有多重含义的情况下。
The PROSLU dataset was co-created by the Research Center for Social Computing and Information Retrieval of Harbin Institute of Technology and Huawei Technologies Co., Ltd. It contains over 5,000 Chinese utterances, each paired with detailed personal profile information including knowledge graphs, user profiles, and context-aware information. The dataset is manually annotated to ensure high quality, and aims to address the problem that traditional text-based spoken language understanding models may fail to accurately recognize intents and slots in real-world scenarios with ambiguous semantics. Its application scenarios mainly focus on enhancing the understanding and response capabilities of dialogue systems in complex environments, especially when user intents are unclear or utterances carry multiple meanings.

- 1Text is no more Enough! A Benchmark for Profile-based Spoken Language Understanding社会计算与信息检索研究中心 · 2022年



