Papers by Yurii Laba

1 papers
From Benchmark to Better Embeddings: Leveraging Synonym Substitution to Enhance Multimodal Models in Ukrainian (2025.findings-emnlp)

Copied to clipboard

Challenge: SSA for Ukrainian degrades more than English in multimodal retrieval . we propose synonym-augmented fine-tuning to mitigate this problem .
Approach: They evaluate dictionary-based and LLM-based methods for Ukrainian multimodal retrieval . they propose a Hybrid method that filters dictionary candidates with an LLM to preserve sense and grammar .
Outcome: The proposed method boosts robustness without harming original performance.

What is GenGO?

GenGO is an NLP powered publication search system. It currenctly indexes 30k+ papers from ACL Anthology, and implements multi-aspect summarization, semantic search, and more!

Information

About
Limitations