A Simple Approach to Case-Based Reasoning in Knowledge Bases
This addresses the challenge of accurate and efficient reasoning in knowledge bases, particularly in low-data scenarios, with potential applications in AI systems requiring interpretable and robust knowledge inference.
The paper tackles the problem of reasoning in knowledge graphs by proposing a non-parametric approach that finds graph path patterns to derive logical rules for queries, achieving new state-of-the-art accuracy on NELL-995 and FB-122 datasets and demonstrating robustness in low data settings.
We present a surprisingly simple yet accurate approach to reasoning in knowledge graphs (KGs) that requires \emph{no training}, and is reminiscent of case-based reasoning in classical artificial intelligence (AI). Consider the task of finding a target entity given a source entity and a binary relation. Our non-parametric approach derives crisp logical rules for each query by finding multiple \textit{graph path patterns} that connect similar source entities through the given relation. Using our method, we obtain new state-of-the-art accuracy, outperforming all previous models, on NELL-995 and FB-122. We also demonstrate that our model is robust in low data settings, outperforming recently proposed meta-learning approaches