Contextual Semantic Parsing using Crowdsourced Spatial Descriptions

TL;DR

This paper introduces a contextual semantic parser that leverages spatial context to improve parsing accuracy of robot commands, achieving significant gains over non-contextual methods by efficiently integrating situational information.

Contribution

It presents a novel approach to semantic parsing that uses spatial context and dynamic programming to disambiguate readings, outperforming traditional non-contextual parsers.

Findings

Achieved 96.53% exact-match accuracy on recognized sentences

Reduced parsing time to near linear by early elimination of incompatible analyses

Demonstrated the effectiveness of spatial context in disambiguating semantic parses

Abstract

We describe a contextual parser for the Robot Commands Treebank, a new crowdsourced resource. In contrast to previous semantic parsers that select the most-probable parse, we consider the different problem of parsing using additional situational context to disambiguate between different readings of a sentence. We show that multiple semantic analyses can be searched using dynamic programming via interaction with a spatial planner, to guide the parsing process. We are able to parse sentences in near linear-time by ruling out analyses early on that are incompatible with spatial context. We report a 34% upper bound on accuracy, as our planner correctly processes spatial context for 3,394 out of 10,000 sentences. However, our parser achieves a 96.53% exact-match score for parsing within the subset of sentences recognized by the planner, compared to 82.14% for a non-contextual parser.