One SPRING to Rule Them Both: Symmetric AMR Semantic Parsing and Generation without a Complex Pipeline

In Text-to-AMR parsing, current state-of-the-art semantic parsers use cumbersome pipelines integrating several different modules or components, and exploit graph recategorization, i.e., a set of content-specific heuristics that are developed on the basis of the training set. However, the generalizability of graph recategorization in an out-of-distribution setting is unclear. In contrast, state-of-the-art AMR-to-Text generation, which can be seen as the inverse to parsing, is based on simpler seq2seq. In this paper, we cast Text-to-AMR and AMR-to-Text as a symmetric transduction task and show that by devising a careful graph linearization and extending a pretrained encoder-decoder model, it is possible to obtain state-of-the-art performances in both tasks using the very same seq2seq approach, i.e., SPRING (\textit{\acl{spring}}). Our model does not require complex pipelines, nor heuristics built on heavy assumptions. In fact, we drop the need for graph recategorization, showing that this technique is actually harmful outside of the standard benchmark. Finally, we outperform the previous state of the art on the English AMR 2.0 dataset by a large margin: on Text-to-AMR we obtain an improvement of 3.6 Smatch points, while on AMR-to-Text we outperform the state of the art by 11.2 BLEU points. We release the software at github.com/SapienzaNLP/spring.

PDF Abstract

Results from the Paper


Task Dataset Model Metric Name Metric Value Global Rank Uses Extra
Training Data
Benchmark
AMR-to-Text Generation Bio SPRING DFS BLEURT 5.2 # 2
AMR-to-Text Generation Bio SPRING DFS + silver BLEURT 5.9 # 1
AMR Parsing Bio SPRING DFS + silver Smatch 59.5 # 4
AMR Parsing Bio SPRING DFS Smatch 59.7 # 3
AMR-to-Text Generation LDC2017T10 SPRING DFS BLEU 45.3 # 4
AMR Parsing LDC2017T10 SPRING Smatch 84.3 # 9
AMR-to-Text Generation LDC2017T10 SPRING DFS + silver BLEU 45.9 # 3
AMR-to-Text Generation LDC2020T02 SPRING DFS BLEU 44.9 # 4
AMR-to-Text Generation LDC2020T02 SPRING DFS + silver BLEU 46.5 # 3
AMR Parsing LDC2020T02 SPRING DFS + silver Smatch 83.0 # 7
AMR Parsing LDC2020T02 SPRING DFS Smatch 83.0 # 7
AMR Parsing New3 SPRING DFS + silver Smatch 71.8 # 4
AMR-to-Text Generation New3 SPRING DFS BLEURT 51.7 # 1
AMR Parsing New3 SPRING DFS Smatch 73.7 # 3
AMR-to-Text Generation The Little Prince SPRING DFS BLEURT 41.5 # 1
AMR Parsing The Little Prince SPRING DFS Smatch 77.3 # 4
AMR Parsing The Little Prince SPRING DFS + silver Smatch 77.5 # 3

Methods