Multilingual Backbone Selection for Zero-Shot Retrieval Performance in Hybrid Batch Training

Information retrieval across different languages is an increasingly important challenge in natural language processing. Recent approaches based on multilingual pre-trained language models have achieved remarkable success, yet they often optimize for either monolingual, cross-lingual, or multilingual retrieval performance at the expense of others. This paper proposes a novel hybrid batch training strategy to simultaneously improve zero-shot retrieval performance across monolingual, cross-lingual, and multilingual settings while mitigating language bias. The approach fine-tunes multilingual lang

Research goal: How does the choice of multilingual pre-trained language model backbone (e.g., mBERT vs. XLM-R vs. mT5) affect the zero-shot retrieval performance of hybrid batch training across typologically diverse languages in MTEB benchmarks?

Autonomous synthesis report generated by Assignee Research. Tribunal consensus score: 8.2/10.