Abstract:In conversational systems, we can use external knowledge to generate more diverse sentences and make these sentences contain actual knowledge. Leveraging knowledge for conversation system is important but challenging. Firstly, the conversation system needs to find the appropriate knowledge. Secondly, the knowledge needs to be coded effectively and generated into fluent utterances. In this paper, we propose a knowledge-driven conversation system to address the above challenges. This system consists of three modules, namely topic predictor, knowledge selector, and dialogue generator. The topic predictor uses a combination of non-deep learning (coarse-grained) and deep learning (fine-grained) to form a rough recall and fine sorting process, and uses them to predict conversation topics. The knowledge selector aims to find the most appropriate knowledge based on the filtered topics. We propose the Bert2Transformer model as our dialogue generator, which can generate rich and fluent utterances based on contextual and relevant knowledge. On the public corpus KdConv, our system outperforms a strong baseline and achieves state-of-the-art results. In the ablation study, we analyze the effectiveness of the proposed components in detail and investigate the performance factors that may affect the knowledge-driven conversation generation. Experimental results show the proposed system achieves a significant improvement compared with traditional baseline methods. The average BLEU scores of our system achieve 35.92 and 23.24, respectively, given appropriate knowledge and without appropriate knowledge, while the Distinct-2 scores of our system achieve 16.32 and 15.93, respectively. The training corpus is publicly available (https://github.com/wulaoshi/dialogue_train_data).

An Ensemble Dialogue System for Facts-Based Sentence Generation

Fact-based Dialogue Generation with Convergent and Divergent Decoding

Prediction, selection, and generation: a knowledge-driven conversation system

Generate, Evaluate, and Select: A Dialogue System with a Response Evaluator for Diversity-Aware Response Generation

Diverse and Informative Dialogue Generation with Context-Specific Commonsense Knowledge Awareness

Generating Diverse Conversation Responses by Creating and Ranking Multiple Candidates

A Knowledge Enhanced Generative Conversational Service Agent

An Ensemble of Retrieval-Based and Generation-Based Human-Computer Conversation Systems

Dynamically Retrieving Knowledge via Query Generation for Informative Dialogue Generation

Learning to Retrieve Entity-Aware Knowledge and Generate Responses with Copy Mechanism for Task-Oriented Dialogue Systems

Two Are Better Than One: an Ensemble of Retrieval- and Generation-Based Dialog Systems.

Emotion-and-knowledge grounded response generation in an open-domain dialogue setting

Detecting Response Generation Not Requiring Factual Judgment

TopicKA: Generating Commonsense Knowledge-Aware Dialogue Responses Towards the Recommended Topic Fact

Topic-Aware Response Generation in Task-Oriented Dialogue with Unstructured Knowledge Access

Unstructured Text Enhanced Open-domain Dialogue System: A Systematic Survey

Improving Knowledge-Aware Dialogue Response Generation by Using Human-Written Prototype Dialogues.

ConfNet2Seq: Full Length Answer Generation from Spoken Questions

Enhancing Generative Conversational Service Agents with Dialog History and External Knowledge

Select, Extend, and Generate: Generative Knowledge Selection for Open-Domain Dialogue Response Generation.

Dual Semantic Knowledge Composed Multimodal Dialog Systems