Archaeology at WE-2026 PARSEME 2.0 Subtask 1 and 2: Parsing is for Encoders, Paraphrasing is for LLMs

Read full text Summarise with VIRAT
Authors: Rares-Alexandru Roscan, Sergiu Nisioi
Cited by: 0
References: 13

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

This paper presents our approach to the PARSEME 2.0 Shared Task on Romanian, covering both Identification (Subtask 1) and Paraphrasing (Subtask 2).While Large Language Models (LLMs) excel at semantic generation, we hypothesize that they lack the structural precision required for MWE identification, leading to "boundary hallucinations" that compromise downstream simplification.Our Rank 1 results on Romanian confirm this: a specialized encoder (RoBERT) using standard sequence labeling outperforms both few-shot LLMs and complex structural parsers (MTLB-STRUCT).This justifies our proposed pipeline: using encoders as precise "pointers" to guide the generative power of LLMs.

Fields of Study & Subjects

Natural Language Processing Techniques Topic Modeling Syntax Semantics Linguistic Variation

Publication details

Journal / source
Workshop on Multiword Expressions (MWE)
Pages
237–247
Publication date
1 January 2026
Language
EN