End-to-End Learning for Answering Structured Queries Directly over Text
This work addresses the challenge of querying text data efficiently for users of structured query languages, though it is incremental as it builds on existing methods like Triple Pattern Fragments and extractive QA.
The paper tackles the problem of answering structured queries directly over text without using a database, focusing on knowledge bases with entities and relations. It achieves an average F1 score of 0.70 across 572 Wikidata relations by combining distributed query answering with extractive question answering models.
Structured queries expressed in languages (such as SQL, SPARQL, or XQuery) offer a convenient and explicit way for users to express their information needs for a number of tasks. In this work, we present an approach to answer these directly over text data without storing results in a database. We specifically look at the case of knowledge bases where queries are over entities and the relations between them. Our approach combines distributed query answering (e.g. Triple Pattern Fragments) with models built for extractive question answering. Importantly, by applying distributed querying answering we are able to simplify the model learning problem. We train models for a large portion (572) of the relations within Wikidata and achieve an average 0.70 F1 measure across all models. We also present a systematic method to construct the necessary training data for this task from knowledge graphs and describe a prototype implementation.