REMI: Mining Intuitive Referring Expressions on Knowledge Bases
This work addresses the need for efficient and intuitive RE mining in applications like natural language generation and data maintenance, representing an incremental improvement in speed over existing methods.
The paper tackles the problem of mining intuitive referring expressions (REs) from large RDF knowledge bases, with results showing that REMI finds user-deemed intuitive REs and is several orders of magnitude faster than an inductive logic programming approach.
A referring expression (RE) is a description that identifies a set of instances unambiguously. Mining REs from data finds applications in natural language generation, algorithmic journalism, and data maintenance. Since there may exist multiple REs for a given set of entities, it is common to focus on the most intuitive ones, i.e., the most concise and informative. In this paper we present REMI, a system that can mine intuitive REs on large RDF knowledge bases. Our experimental evaluation shows that REMI finds REs deemed intuitive by users. Moreover we show that REMI is several orders of magnitude faster than an approach based on inductive logic programming.