Evaluating Semantic Parsing against a Simple Web-based Question Answering Model

TL;DR

This paper compares semantic parsing-based question answering models to a simple web-based baseline on complex questions, revealing strengths in conjunctions but challenges with relation composition and superlatives.

Contribution

It introduces a novel evaluation approach for semantic parsing QA models against a web-based baseline on the COMPLEXQUESTIONS dataset.

Findings

Semantic parsing models achieve 35 F1, close to 41 of state-of-the-art.

Models perform well on conjunctions in complex questions.

Models struggle with relation composition and superlatives.

Abstract

Semantic parsing shines at analyzing complex natural language that involves composition and computation over multiple pieces of evidence. However, datasets for semantic parsing contain many factoid questions that can be answered from a single web document. In this paper, we propose to evaluate semantic parsing-based question answering models by comparing them to a question answering baseline that queries the web and extracts the answer only from web snippets, without access to the target knowledge-base. We investigate this approach on COMPLEXQUESTIONS, a dataset designed to focus on compositional language, and find that our model obtains reasonable performance (35 F1 compared to 41 F1 of state-of-the-art). We find in our analysis that our model performs well on complex questions involving conjunctions, but struggles on questions that involve relation composition and superlatives.