Reinforcement Learning Over Knowledge Graphs for Explainable Dialogue Intent Mining

Kai Yang,Jie Zhang,Yafang Wang,Gerard De Melo,Xinyu Kong

doi:10.1109/access.2020.2991257

Kai Yang, Jie Zhang + Show 3 more

Open Access

https://doi.org/10.1109/access.2020.2991257

Copy DOI

Abstract

In light of the millions of households that have adopted intelligent assistant powered devices, multi-turn dialogue has become an important field of inquiry. Most current methods identify the underlying intent in the dialogue using opaque classification techniques that fail to provide any interpretable basis for the classification. To address this, we propose a scheme to interpret the intent in multi-turn dialogue based on specific characteristics of the dialogue text. We rely on policy-guided reinforcement learning to identify paths in a graph to confirm concrete paths of inference that serve as interpretable explanations. The graph is induced based on the multi-turn dialogue user utterances, the intents, i.e., standard queries of the dialogues, and the sub-intents associated with the dialogues. Our reinforcement learning method then discerns the characteristics of the dialogue in chronological order as the basis for multi-turn dialogue path selection. Finally, we consider a wide range of recently proposed knowledge graph-based recommender systems as baselines, mostly based on deep reinforcement learning and our method performs best.

Highlights

IntroductionMillions of households have adopted intelligent assistant powered devices
Across the globe, millions of households have adopted intelligent assistant powered devices
We propose a method called PGMD that draws on a neural reinforcement learning network to navigate the knowledge graph in pursuit of the pertinent query nodes in the graph

Summary

Introduction

Millions of households have adopted intelligent assistant powered devices. The system needs to identify a user’s information need from this dialogue and locate an appropriate answer from all the knowledge that is accessible to it Such knowledge can oftentimes be regarded as taking the form of a knowledge graph and locating an answer often corresponds to identifying relevant nodes in the graph [1]. Recent work in this area has exploited advances in neural representation learning to address this task [2], [3]. In real-world deployments of such systems, it is not sufficient for a multi-turn dialogue recognition system to merely use latent vector representations for knowledge graph nodes to identify appropriate responses.

Objectives

Results

Conclusion