Xiaoice Firm introduced the official launch of its new “Zero-Shot” digital human expertise (Zero-shot Xiaoice Neural Rendering, Zero-XNR) and concurrently launched a brand new, universally accessible digital worker product primarily based on Z-XNR expertise, additional enriching the Xiaoice AI digital worker product line.
Zero-Shot Know-how and Massive Mannequin Integration
The brand new expertise adheres to the “Digital Human + Massive Mannequin” method that the Xiaoice AI digital worker product line constantly follows. In comparison with different current applied sciences, this new development leverages a trillion-parameter mannequin base and digital human interplay suites constructed on giant fashions. It compresses the coaching information required for digital people to “seconds,” making customization nearly instantaneous. The generated digital people can be utilized immediately for real-time interplay, a worldwide first. The brand new expertise maintains the ultra-high-definition requirements of Xiaoice AI digital worker merchandise, reaching lifelike recreations of human workers’ appearances and voices.
Key Upgrades to Xiaoice Digital Human Know-how
- New Z-XNR Know-how
The fusion framework of Z-XNR “Zero-Shot” digital human expertise and TTS voice giant fashions permits for high-quality picture and voice replication inside seconds. The distinctive information coaching and inference algorithms guarantee high-quality replication, customized visible expression, and rapid availability. - New Trillion-Parameter Mannequin Base and Agent Building Framework
Based mostly on the trillion-parameter mannequin base, the upgraded hybrid base structure’s Agent building framework builds a complete platform for creating and enhancing digital worker capabilities—the Digital Mind Platform. This equips your complete collection of digital workers with highly effective and various skilled interplay suites, enabling them to raised perceive prospects and enterprise data and enterprise, with interactions exactly tailor-made to native enterprise workflows. - New Enhanced Audio-Visible Transmission System
Using a self-developed engine primarily based on WebRTC, it offers ultra-high-definition video streaming capabilities, with telecom-grade high-quality, low-latency audio-visual transmission and strong community packet loss resistance. It ensures exact audio-visual synchronization and may be built-in with…
This textual content offers an outline of the most recent technological developments in Xiaoice’s AI digital worker options, highlighting important enhancements in real-time digital human interactions and enhanced communication capabilities.