Sensory enables ChatGPT to give human-like voice responses on consumer devices
Sensory has introduced a new technology integration that allows ChatGPT and other Large Language Models to provide conversational voice responses on devices without keyboards or big screens.
The company says Sensory’s technology can enable consumer devices to generate human-like voice responses, creating a new generation of voice assistants that can be customized for different domains.
Sensory’s technology integration enables a conversational experience on consumer electronics, allowing for voice chat-type capabilities. It can be used in-ear voice assistants, smartwatches, smartphones and automotive infotainment systems.
Sensory’s voice AI stack provides users with wake word recognition, speech-to-text with context, intelligent response selection and text-to-speech capabilities. It also enables natural human-like conversations by allowing follow-up questions and commands to filter or add more information.
The hybrid cloud and edge AI platform from Sensory allows customers to choose which AI technologies they want to use, splitting the tasks between edge devices and the cloud. Light-duty AI can run on-device, like wake word recognition and voice controls.
At the same time, more complex inferencing models such as natural language understanding or voice chat can route to a connected, powerful device such as a smartphone. High-horsepower AI, such as face and object recognition can route to the cloud for improved revalidation.
SensoryCloud’s voice assistant solution utilizes Go, gRPC, NVIDIA Triton, and AWS Global Accelerator to build highly-scalable applications that can handle complex workloads. Dialog data is compressed using proprietary techniques to minimize cloud fees and reduce latency.
Article Topics
AI/ML | AWS | ChatGPT | conversational voice | generative AI | GPU | Inc. | IoT | Nvidia | Sensory | voice recognition
Comments