Speech-to-Speech Model Research at Google DeepMind — Valeria Wu Fon & Tom Ouyang, Google DeepMind
Valeria Wu Fon and Tom Ouyang from Google DeepMind present research on speech-to-speech models — systems that listen to speech and respond with speech directly, without converting to text.
Before 2018, recognizing speech required multiple separate components; modern models do it end-to-end. Their model trains on audio, video, and text together, allowing it to handle things nobody explicitly programmed — for example, keeping English terms that a Spanish speaker actually uses. They identify three tensions: the system must be fast for natural conversation, intelligent to understand instructions, and multimodal to accept both speech and visual material. Increasing one capability often weakens the others, and they demonstrate this with demos of live translation in meetings, roadside assistance for vehicle registration, and the system's ability to know when it's not its turn to speak.
Vibekollen prepared this summary with AI from the original publication. The content belongs to AI Engineer.
More to read
Lumaril, which develops AI software for optimizing industrial production by connecting to existing control systems, raised a $66M Series A at a $400M valuation
Techmeme 4 h ago
DeepSeek open sources Super-Agent as a research prototype, a proactive coding agent it says can monitor group chats for coding issues and suggest solutions
Techmeme 5 h ago
Insurance software maker Guidewire agrees to acquire Hyperexponential, which makes AI-powered insurance underwriting and pricing software, for $1.7B in cash
Techmeme 5 h ago
One word for AI agents in 2027
Nate Herk 9 h ago