Predicting the Target Word of Game-playing Conversations using a Low-Rank Dialect Adapter for Decoder Models

TL;DR

This paper introduces LoRDD, a low-rank dialect adapter for decoder models that improves target word prediction in dialectal conversations, reducing performance gaps across dialects in game-playing dialogue datasets.

Contribution

It extends dialect adapters to decoder models using contrastive learning, demonstrating improved dialect adaptation in target word prediction tasks.

Findings

LoRDD outperforms four baselines on Indian and Nigerian English conversations.

It reduces the dialect performance gap significantly.

The approach is effective for dialect adaptation in decoder models.

Abstract

Dialect adapters that improve the performance of LLMs for NLU tasks on certain sociolects/dialects/national varieties ('dialects' for the sake of brevity) have been reported for encoder models. In this paper, we extend the idea of dialect adapters to decoder models in our architecture called LoRDD. Using MD-3, a publicly available dataset of word game-playing conversations between dialectal speakers, our task is Target Word Prediction (TWP) from a masked conversation. LoRDD combines task adapters and dialect adapters where the latter employ contrastive learning on pseudo-parallel conversations from MD-3. Our experiments on Indian English and Nigerian English conversations with two models (Mistral and Gemma) demonstrate that LoRDD outperforms four baselines on TWP. Additionally, it significantly reduces the performance gap with American English, narrowing it to 12% and 5.8% for word similarity, and 25% and 4.5% for accuracy, respectively. The focused contribution of LoRDD is in its promise for dialect adaptation of decoder models using TWP, a simplified version of the commonly used next-word prediction task.