SOKE introduces a novel approach for translating text into sign language across multiple languages, addressing a critical gap in accessibility technology.

• Leverages a retrieval-enhanced framework that treats signs as discrete tokens
• Supports multilingual generation of sign language from text input
• Achieves state-of-the-art performance across multiple metrics and datasets
• Enables practical applications for deaf and hard-of-hearing communities

This research represents a significant advancement for linguists and accessibility specialists by bridging communication gaps between signed and spoken languages, potentially revolutionizing how technology can support sign language users worldwide.

Original Paper: Signs as Tokens: A Retrieval-Enhanced Multilingual Sign Language Generator