SOTAVerified

ÚFAL at MRP 2020: Permutation-invariant Semantic Parsing in PERIN

2020-11-02Code Available1· sign in to hype

David Samuel, Milan Straka

Code Available — Be the first to reproduce this paper.

Reproduce

Code

Abstract

We present PERIN, a novel permutation-invariant approach to sentence-to-graph semantic parsing. PERIN is a versatile, cross-framework and language independent architecture for universal modeling of semantic structures. Our system participated in the CoNLL 2020 shared task, Cross-Framework Meaning Representation Parsing (MRP 2020), where it was evaluated on five different frameworks (AMR, DRG, EDS, PTG and UCCA) across four languages. PERIN was one of the winners of the shared task. The source code and pretrained models are available at https://github.com/ufal/perin.

Tasks

Benchmark Results

DatasetModelMetricClaimedVerifiedStatus
AMR (chinese, MRP 2020)PERINF180.52Unverified
AMR (english, MRP 2020)PERINF180.23Unverified
DRG (english, MRP 2020)PERINF194.16Unverified
DRG (german, MRP 2020)PERINF189.83Unverified
EDS (english, MRP 2020)PERINF192.73Unverified
PTG (czech, MRP 2020)PERINF192.24Unverified
PTG (english, MRP 2020)PERINF189.19Unverified
UCCA (english, MRP 2020)PERINF176.4Unverified
UCCA (german, MRP 2020)PERINF181.01Unverified

Reproductions