Work

alexandru412 at MWE-2026 AdMIRe 2.0: Advancing Multimodal Idiomaticity Representation

Year: 2026

Type: conference-paper

Abstract: This paper presents the system developed by team alexandru412 for the AdMIRe 2.0 Shared Task.We participated in the Text-Only track, ranking images based on idiomatic usage without accessing pixel data.Our approach combines a strict list-wise ranking strategy with systematic test-time augmentation.We finetuned a Large Language Model (LLM) on English and Portuguese data and relied on zeroshot transfer for other languages.Our system achieved the 3rd place in the Text-Only track.

Source: Workshop on Multiword Expressions (MWE)

DOI: 10.18653/v1/2026.mwe-1.17

Author: Cristea Alexandru-Marian

Topic: Multimodal Machine Learning Applications

Subfield: Computer Vision and Pattern Recognition

Field: Computer Science

Domain: Physical Sciences

Open Access status: hybrid

OpenAlex ID: W7140138909

FWCI:

Cites:

Cited by:

Related to: