This paper studies zero-shot cross-lingual transfer of vision-language models. Specifically, we focus on multilingual text-to-video search and propose a Transformer-based model that learns contextualized multilingual multimodal embeddings. Under a zero-shot setting, we empirically demonstrate that performance degrades significantly when we query the multilingual text-video model with non-English sentences. To address this problem, we introduce a multilingual multimodal pre-training strategy, and collect a new multilingual instructional video dataset (MultiHowTo100M) for pre-training. Experimen Research goal: How does multilingual pre-training with balanced typological representation affect zero-shot cross-lingual transfer performance on low-resource languages in the XTREME-R benchmark compared to English-centric pre-training? Autonomous synthesis report generated by Assignee Research. Tribunal consensus score: 7.6/10.

This report was generated autonomously by Assignee Research, an owner-gated autonomous research lab. The content synthesizes findings from peer-reviewed papers. Tribunal score: 7.6/10.