NMIXX: Domain-Adapted Neural Embeddings for Cross-Lingual eXploration of Finance

TL;DR

This paper introduces NMIXX, a domain-adapted cross-lingual embedding model for finance, and KorFinSTS, a Korean financial semantic similarity benchmark, improving multilingual financial understanding especially in low-resource languages.

Contribution

We develop NMIXX, a novel domain-specific multilingual embedding model fine-tuned with specialized triplets, and release KorFinSTS, a benchmark for Korean financial semantic similarity evaluation.

Findings

NMIXX outperforms baseline models on financial semantic tasks.

Rich Korean token coverage enhances low-resource language adaptation.

Models and benchmark are publicly available for community use.

Abstract

General-purpose sentence embedding models often struggle to capture specialized financial semantics, especially in low-resource languages like Korean, due to domain-specific jargon, temporal meaning shifts, and misaligned bilingual vocabularies. To address these gaps, we introduce NMIXX (Neural eMbeddings for Cross-lingual eXploration of Finance), a suite of cross-lingual embedding models fine-tuned with 18.8K high-confidence triplets that pair in-domain paraphrases, hard negatives derived from a semantic-shift typology, and exact Korean-English translations. Concurrently, we release KorFinSTS, a 1,921-pair Korean financial STS benchmark spanning news, disclosures, research reports, and regulations, designed to expose nuances that general benchmarks miss. When evaluated against seven open-license baselines, NMIXX's multilingual bge-m3 variant achieves Spearman's rho gains of +0.10 on English FinSTS and +0.22 on KorFinSTS, outperforming its pre-adaptation checkpoint and surpassing other models by the largest margin, while revealing a modest trade-off in general STS performance. Our analysis further shows that models with richer Korean token coverage adapt more effectively, underscoring the importance of tokenizer design in low-resource, cross-lingual settings. By making both models and the benchmark publicly available, we provide the community with robust tools for domain-adapted, multilingual representation learning in finance.