使用 Amazon Nova Sonic 创建语音优先应用程序.pdf

1、A I M 3 7 7Osman IpekApplied AI ArchitectAmazon AGIReilly MantonSolutions ArchitectAWS TelecomCreate voice-first applications with Amazon Nova Sonic 2025,Amazon Web Services,Inc.or its affiliates.All rights reserved.What if having human-like interactionswith an AIwas aseasyas calling one single API?

2、Agenda Nova Sonic refresher Demo Interactive whiteboarding:building a Nova Sonic agent Q&A ConclusionAmazon Nova SonicState-of-the-art speech-to-speech model that delivers real-time,human-like voice-first conversationsReal-time streamingBidirectional streaming Bedrock API inregions IAD,NRT,ARN,and P

3、DX(new!)ModalitiesInput:Speech and text(new!)Output:Speech and textContext window1 million tokensLanguagesEnglish,French,Italian,German,Spanish,Portuguese,and HindiVoices16 voices in total,including polyglot voicesand code-switchingControllability featureTurn-taking-length of pause detectionAgentic

4、featuresTool use for function calling,asynchronous/background task completion,knowledge grounding Best-in-class speech recognition:Accurately recognizes streaming speech across accents with robustness to background noise Fluid dialog handling and natural turn taking:Handles users speech interruption

5、s,detects non-verbal cues(e.g.laughter,grunts,inter-sentential pause and hesitation)to enable human like turn taking in dialogs Adaptive speech responses:Adapts speech response to users tone and sentiment.Conversational AI featuresBuilding on top of what weve had since April with Nova Sonic 1.0 2025

6、,Amazon Web Services,Inc.or its affiliates.All rights reserved.Demo 2025,Amazon Web Services,Inc.or its affiliates.All rights reserved.Quick poll Please raise your hand if youreBuilding for contact centers?Integrating with existing phone systems?Building web/mobile voice assistants?In prototyping ph