OpenDialKG: Explainable Conversational Reasoning with Attention-based Walks over Knowledge Graphs

TLDR

The study proposes the DialKG Walker model, a conversational reasoning system that learns symbolic transitions over a large knowledge graph to predict natural, contextually diverse entities using an attention‑based graph path decoder. The authors constructed the OpenDialKG corpus of 15,000 human‑to‑human dialogs annotated with entity and path references from a 1‑million‑fact knowledge graph, and built the DialKG Walker model to learn symbolic transitions as structured traversals over the KG. Evaluation demonstrates that DialKG Walker produces more natural, human‑like responses than state‑of‑the‑art baselines across in‑domain and cross‑domain tasks, and it generates interpretable KG walk paths that explain the conversational reasoning.

Abstract

We study a conversational reasoning model that strategically traverses through a large-scale common fact knowledge graph (KG) to introduce engaging and contextually diverse entities and attributes. For this study, we collect a new Open-ended Dialog <-> KG parallel corpus called OpenDialKG, where each utterance from 15K human-to-human role-playing dialogs is manually annotated with ground-truth reference to corresponding entities and paths from a large-scale KG with 1M+ facts. We then propose the DialKG Walker model that learns the symbolic transitions of dialog contexts as structured traversals over KG, and predicts natural entities to introduce given previous dialog contexts via a novel domain-agnostic, attention-based graph path decoder. Automatic and human evaluations show that our model can retrieve more natural and human-like responses than the state-of-the-art baselines or rule-based models, in both in-domain and cross-domain tasks. The proposed model also generates a KG walk path for each entity retrieved, providing a natural way to explain conversational reasoning.