A Knowledge-Grounded Multimodal Search-Based Conversational Agent
A Knowledge-Grounded Multimodal Search-Based Conversational Agent
复制标题
DOI：
10.18653/v1/w18-5709
复制
发表时间：
2018-10
期刊：
影响因子：
--
通讯作者：
Shubham Agarwal;Ondrej Dusek;Ioannis Konstas;Verena Rieser
中科院分区：
文献类型：
--
作者：
Shubham Agarwal;Ondrej Dusek;Ioannis Konstas;Verena Rieser
基于多模式搜索的对话是一项具有挑战性的新任务：它将基于视觉的问答系统扩展为可访问外部数据库的多轮对话。我们通过从最近发布的多模式对话(MMD)数据集(Saha等人，2017年)学习神经反应生成系统来应对这一新挑战。我们介绍了一种基于知识的多通道对话模型，其中编码的知识库(KB)表示被附加到解码器输入。在基于文本的相似性度量方面，我们的模型大大优于强基线(超过9个BLEU点，其中3个完全是由于使用了知识库中的额外信息)。
Multimodal search-based dialogue is a challenging new task: It extends visually grounded question answering systems into multi-turn conversations with access to an external database. We address this new challenge by learning a neural response generation system from the recently released Multimodal Dialogue (MMD) dataset (Saha et al., 2017). We introduce a knowledge-grounded multimodal conversational model where an encoded knowledge base (KB) representation is appended to the decoder input. Our model substantially outperforms strong baselines in terms of text-based similarity measures (over 9 BLEU points, 3 of which are solely due to the use of additional information from the KB).