Few-Shot Semantic Parsing for New Predicates

TL;DR

This paper introduces a novel few-shot semantic parsing approach that leverages meta-learning, attention regularization, and smoothing techniques, significantly improving accuracy on benchmarks with minimal training examples.

Contribution

It presents a new method combining meta-learning, attention regularization, and smoothing to enhance few-shot semantic parsing performance.

Findings

Outperforms baseline models in one-shot and two-shot settings

Achieves over 25% accuracy on benchmark datasets with minimal examples

Demonstrates effectiveness of combined techniques in low-resource semantic parsing

Abstract

In this work, we investigate the problems of semantic parsing in a few-shot learning setting. In this setting, we are provided with utterance-logical form pairs per new predicate. The state-of-the-art neural semantic parsers achieve less than 25% accuracy on benchmark datasets when k= 1. To tackle this problem, we proposed to i) apply a designated meta-learning method to train the model; ii) regularize attention scores with alignment statistics; iii) apply a smoothing technique in pre-training. As a result, our method consistently outperforms all the baselines in both one and two-shot settings.