XLM-R Domain Adaptation: English vs. Multilingual Intermediate-Task Training Effects

Transfer learning from large language models (LLMs) has emerged as a powerful technique to enable knowledge-based fine-tuning for a number of tasks, adaptation of models for different domains and even languages. However, it remains an open question, if and when transfer learning will work, i.e. leading to positive or negative transfer. In this paper, we analyze the knowledge transfer across three natural language processing (NLP) tasks - text classification, sentimental analysis, and sentence similarity, using three LLMs - BERT, RoBERTa, and XLNet - and analyzing their performance, by fine-tun

Research goal: How does the robustness of XLM-R to domain shifts in cross-lingual tasks compare when using English intermediate-task training versus multilingual intermediate-task training, measured by accuracy on domain-specific subsets of the XTREME-R benchmark?

Autonomous synthesis report generated by Assignee Research. Tribunal consensus score: 8.5/10.