ConCSE: Unified Contrastive Learning and Augmentation for Code-Switched Embeddings

TL;DR

This paper introduces ConCSE, a novel contrastive learning method for code-switched embeddings, validated on a new English-Korean dataset, demonstrating improved semantic similarity performance.

Contribution

It presents a new unified contrastive learning and augmentation approach specifically designed for code-switched language embeddings, along with a new dataset for English-Korean CS scenarios.

Findings

ConCSE improves semantic similarity scores by 1.77% on Koglish-STS.

The Koglish dataset highlights the need for CS-specific resources.

Multilingual models show differential performance on monolingual versus CS data.

Abstract

This paper examines the Code-Switching (CS) phenomenon where two languages intertwine within a single utterance. There exists a noticeable need for research on the CS between English and Korean. We highlight that the current Equivalence Constraint (EC) theory for CS in other languages may only partially capture English-Korean CS complexities due to the intrinsic grammatical differences between the languages. We introduce a novel Koglish dataset tailored for English-Korean CS scenarios to mitigate such challenges. First, we constructed the Koglish-GLUE dataset to demonstrate the importance and need for CS datasets in various tasks. We found the differential outcomes of various foundation multilingual language models when trained on a monolingual versus a CS dataset. Motivated by this, we hypothesized that SimCSE, which has shown strengths in monolingual sentence embedding, would have limitations in CS scenarios. We construct a novel Koglish-NLI (Natural Language Inference) dataset using a CS augmentation-based approach to verify this. From this CS-augmented dataset Koglish-NLI, we propose a unified contrastive learning and augmentation method for code-switched embeddings, ConCSE, highlighting the semantics of CS sentences. Experimental results validate the proposed ConCSE with an average performance enhancement of 1.77\% on the Koglish-STS(Semantic Textual Similarity) tasks.