Bandwidth-Efficient and Privacy-Preserving Edge-Cloud Many-to-Many Speech Translation 事件

Bandwidth-Efficient and Privacy-Preserving Edge-Cloud Many-to-Many Speech Translation arXiv:2605.28642v1 Announce Type: new Abstract: Multimodal large language models (MLLMs) have demonstrated significant potential for speech-to-text translation (S2TT). However, existing deployment paradigms face critical challenges: pure on-device models suffer from resource constraints, while centralized cloud systems incur severe privacy risks and bandwidth bottlenecks by transmitting raw voice data. Further