Heterogeneous Relational Reasoning in Knowledge Graphs with Reinforcement Learning
This work addresses path-based reasoning problems for applications like question answering and recommender systems, representing an incremental improvement over existing RL methods.
The paper tackles the challenges of large action space and accurate neighborhood representation in reinforcement learning for path-based relational reasoning over knowledge graphs by introducing a type-enhanced RL agent that uses GNNs and entity types. The method outperforms state-of-the-art RL methods on a real-world dataset and discovers more novel paths during training.
Path-based relational reasoning over knowledge graphs has become increasingly popular due to a variety of downstream applications such as question answering in dialogue systems, fact prediction, and recommender systems. In recent years, reinforcement learning (RL) has provided solutions that are more interpretable and explainable than other deep learning models. However, these solutions still face several challenges, including large action space for the RL agent and accurate representation of entity neighborhood structure. We address these problems by introducing a type-enhanced RL agent that uses the local neighborhood information for efficient path-based reasoning over knowledge graphs. Our solution uses graph neural network (GNN) for encoding the neighborhood information and utilizes entity types to prune the action space. Experiments on real-world dataset show that our method outperforms state-of-the-art RL methods and discovers more novel paths during the training procedure.