Impact of Batch Composition Ratio on Multilingual Zero-Shot Retrieval Performance in CLIP 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 adjusting the batch composition ratio of monolingual to cross-lingual data in CLIP training impact the performance of multilingual zero-shot image-text retrieval on the XTD benchmark when evaluated with mAP@100 across diverse language families?

Autonomous synthesis report generated by Assignee Research. Tribunal consensus score: 8.1/10.