Papers › Maximum Bayes Smatch Ensemble Distillation for AMR Parsing

Maximum Bayes Smatch Ensemble Distillation for AMR Parsing

14 Dec 2021NAACL 2022 7arXiv:2112.07790archive 2025-07-28

Young-suk Lee, Ramon Fernandez Astudillo, Thanh Lam Hoang, Tahira Naseem, Radu Florian, Salim Roukos

AMR parsing has experienced an unprecendented increase in performance in the last three years, due to a mixture of effects including architecture improvements and transfer learning. Self-learning techniques have also played a role in pushing performance forward. However, for most recent high performant parsers, the effect of self-learning and silver data augmentation seems to be fading. In this paper we propose to overcome this diminishing returns of silver data by combining Smatch-based ensembling techniques with ensemble distillation. In an extensive experimental setup, we push single model English parser performance to a new state-of-the-art, 85.9 (AMR2.0) and 84.3 (AMR3.0), and return to substantial gains from silver data augmentation. We also attain a new state-of-the-art for cross-lingual AMR parsing for Chinese, German, Italian and Spanish. Finally we explore the impact of the proposed technique on domain adaptation, and show that it can produce gains rivaling those of human annotated data for QALD-9 and achieve a new state-of-the-art for BioAMR.

PaperPDFConference PDFCode

In Syntology Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

Code

IBM/transition-amr-parser officialmentioned in papermentioned on GitHubpytorch report
ibm/amr-annotations officialmentioned in paper report
pournaki/transition-amr-parser mentioned on GitHubpytorch report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

Not run by Syntology. Nothing on this page verifies that the listed code works.

Tasks

AMR ParsingData AugmentationDomain AdaptationSelf-LearningTransfer Learning

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
AMR Parsing Bio StructBART + MBSE (IBM) Smatch 66.9 #1 of 5 Archive leaderboard report
AMR Parsing LDC2017T10 StructBART + MBSE (IBM) Smatch 86.7 #1 of 27 Archive leaderboard report
AMR Parsing LDC2017T10 StructBART + MBSE (IBM) Smatch 85.9 #4 of 27 Archive leaderboard report
AMR Parsing LDC2020T02 Graphene Smatch (MBSE paper) (IBM) Smatch 85.4 #1 of 13 Archive leaderboard report
AMR Parsing LDC2020T02 StructBART + MBSE (IBM) Smatch 84.3 #5 of 13 Archive leaderboard report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Methods

Self-Learning

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections