AMRs Assemble! Learning to Ensemble with Autoregressive Models for AMR Parsing

TL;DR

This paper introduces novel ensemble strategies using Transformer models for AMR parsing that improve structural robustness and computational efficiency, addressing limitations of current ensemble methods and metric exploitation.

Contribution

It proposes two new Transformer-based ensemble methods for AMR parsing that enhance structural validity and reduce computation, advancing the state-of-the-art.

Findings

Improved robustness to AMR structural constraints

Reduced computational time for ensemble predictions

Enhanced accuracy by addressing SMATCH metric weaknesses

Abstract

In this paper, we examine the current state-of-the-art in AMR parsing, which relies on ensemble strategies by merging multiple graph predictions. Our analysis reveals that the present models often violate AMR structural constraints. To address this issue, we develop a validation method, and show how ensemble models can exploit SMATCH metric weaknesses to obtain higher scores, but sometimes result in corrupted graphs. Additionally, we highlight the demanding need to compute the SMATCH score among all possible predictions. To overcome these challenges, we propose two novel ensemble strategies based on Transformer models, improving robustness to structural constraints, while also reducing the computational time. Our methods provide new insights for enhancing AMR parsers and metrics. Our code is available at \href{https://www.github.com/babelscape/AMRs-Assemble}{github.com/babelscape/AMRs-Assemble}.