Abstract:Dialogues systems endow machines with the ability to converse with humans using natural language. Nonetheless, previous Seq2Seq-based generative dialogue systems often generate safe but meaningless responses, such as ‘I don't know’ or ‘I think so’. To this end, researchers proposed to infuse external knowledge into dialogue generation, and such knowledge-enhanced methods have achieved remarkable improvements in the open-domain dialogue systems. External knowledge is an exogenous input, where the estrangement inevitably exists between knowledge and dialogue context. Although previous knowledge-enhanced works can already use commonsense knowledge to generate informative responses, they always use knowledge in a single-channel paradigm, which is hard to accurately handle different data-flows and then tends to generate irrational dialogue responses. Thus, they tend to be confused and generate strange responses when infusing the knowledge into dialogue generation, such as ‘I just ate a basketball,’ dramatically degrading the user experience. To address this problem, this paper proposes a novel Channel-Aware Knowledge Fusing Network (CAKF). Rather than following the traditional single-channel paradigm, CAKF employs three unique channels to handle different data-flows more clearly and rationally: a base channel serves like a vanilla Seq2Seq decoder; a context channel to utilize the contextual information, and a knowledge channel to infuse commonsense knowledge into the dialogue generation. Above such three channels, a Sequential Manager is built to maintain the global sequential decision state, aggregate the local data-flows, and make the final prediction. Experiments on two open-released datasets (a Chinese Weibo and an English Reddit) demonstrated the superior performance of this work against various state-of-the-art approaches.

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

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

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

Read Key Points: Dialogue-Grounded Knowledge Points Generation with Multi-Level Salience-Aware Mixture

Topic-level knowledge sub-graphs for multi-turn dialogue generation

Generating Rational Commonsense Knowledge-Aware Dialogue Responses with Channel-Aware Knowledge Fusing Network

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

Knowledge Augmented Dialogue Generation with Divergent Facts Selection

Dynamically Retrieving Knowledge via Query Generation for Informative Dialogue Generation

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

Improving Empathetic Dialogue Generation by Dynamically Infusing Commonsense Knowledge.

Knowledge-aware Dialogue Generation with Hybrid Attention (Student Abstract)

Generative commonsense knowledge subgraph retrieval for open-domain dialogue response generation

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

Knowledge Bridging for Empathetic Dialogue Generation

Coarse-to-Careful: Seeking Semantic-related Knowledge for Open-domain Commonsense Question Answering

Generative Knowledge Selection for Knowledge-Grounded Dialogues

Commonsense Knowledge Aware Conversation Generation with Graph Attention

Commonsense Generative Model for Chinese Automatic Knowledge Graph Construction

More is Better: Enhancing Open-Domain Dialogue Generation via Multi-Source Heterogeneous Knowledge

Think Before You Speak: Explicitly Generating Implicit Commonsense Knowledge for Response Generation