Transition-Based Generation from Abstract Meaning Representations

TL;DR

This paper presents a transition-based method for generating English sentences from Abstract Meaning Representation graphs by transforming them into syntactic structures and applying learned actions, achieving state-of-the-art BLEU scores.

Contribution

It introduces a novel transition-based approach that converts AMR graphs into dependency-like trees for improved sentence generation.

Findings

Achieved a BLEU score of 27.4 on LDC2014T12 test set.

First to report such results without silver standard annotations.

Demonstrated efficiency of the proposed action sequence approximation algorithm.

Abstract

This work addresses the task of generating English sentences from Abstract Meaning Representation (AMR) graphs. To cope with this task, we transform each input AMR graph into a structure similar to a dependency tree and annotate it with syntactic information by applying various predefined actions to it. Subsequently, a sentence is obtained from this tree structure by visiting its nodes in a specific order. We train maximum entropy models to estimate the probability of each individual action and devise an algorithm that efficiently approximates the best sequence of actions to be applied. Using a substandard language model, our generator achieves a Bleu score of 27.4 on the LDC2014T12 test set, the best result reported so far without using silver standard annotations from another corpus as additional training data.