Stochastic Answer Networks for Natural Language Inference

TL;DR

This paper introduces a stochastic answer network that iteratively refines predictions for natural language inference, achieving state-of-the-art results on multiple benchmarks.

Contribution

The paper presents a novel stochastic answer network that employs multi-step inference strategies for improved natural language inference performance.

Findings

SAN achieves state-of-the-art results on SNLI, MultiNLI, and Quora datasets.

Iterative refinement improves inference accuracy.

Model outperforms previous methods on benchmark datasets.

Abstract

We propose a stochastic answer network (SAN) to explore multi-step inference strategies in Natural Language Inference. Rather than directly predicting the results given the inputs, the model maintains a state and iteratively refines its predictions. Our experiments show that SAN achieves the state-of-the-art results on three benchmarks: Stanford Natural Language Inference (SNLI) dataset, MultiGenre Natural Language Inference (MultiNLI) dataset and Quora Question Pairs dataset.