A Fast Unified Model for Parsing and Sentence Understanding

TL;DR

This paper introduces SPINN, a unified neural model that combines parsing and sentence understanding, enabling fast, batched processing and effective handling of unparsed data for large-scale NLP tasks.

Contribution

The paper presents SPINN, a novel model that integrates parsing and interpretation in a single framework, supporting batched computation and unparsed data processing.

Findings

Supports up to 25x speedup over other tree models

Operates effectively on unparsed data with minimal accuracy loss

Outperforms other sentence-encoding models on Stanford NLI task

Abstract

Tree-structured neural networks exploit valuable syntactic parse information as they interpret the meanings of sentences. However, they suffer from two key technical problems that make them slow and unwieldy for large-scale NLP tasks: they usually operate on parsed sentences and they do not directly support batched computation. We address these issues by introducing the Stack-augmented Parser-Interpreter Neural Network (SPINN), which combines parsing and interpretation within a single tree-sequence hybrid model by integrating tree-structured sentence interpretation into the linear sequential structure of a shift-reduce parser. Our model supports batched computation for a speedup of up to 25 times over other tree-structured models, and its integrated parser can operate on unparsed data with little loss in accuracy. We evaluate it on the Stanford NLI entailment task and show that it significantly outperforms other sentence-encoding models.