Sensory, a Silicon Valley innovator of machine learning solutions for speech recognition and biometric identification, has announced the beta release of SensoryCloud.ai, a complete AI as a Service platform designed for processing voice and vision AI workloads in the cloud.
The company’s technologies are widely deployed in consumer electronics applications including mobile phones, automotive, wearables, toys, IoT, PCs, medical products, and various home electronics.
The SensoryCloud platform, which leverages Sensory's experience with voice and vision AI, is launched with AI services such as Speech to Text, Sound Identification, Wake Word Verification, Face Verification, and Speaker Identification. Additional services and updates will be offered throughout the year.
Using AI to help address customer needs
Founded in 1994, Sensory historically focused on AI on the edge. However, many embedded clients indicated a strong desire for cloud solutions with degrees of freedom not available from the typical cloud service providers. "We have a history of building fast and accurate AI models, and we paired this capability with some of the brightest and freshest minds in the cloud industry," states Todd Mozer, Sensory CEO. "The result is a hybrid cloud platform that uses state-of-the-art AI to address customers unique needs for control, flexibility, cost, accuracy, reliability, features, latency, and privacy."
SensoryCloud AI as a Service Includes:
- SensoryCloud Speech to Text (STT) – A world-class GPU-accelerated speech recognition engine that can be quickly customised to process application-specific jargon. Ideal for either streaming or batch-mode operations with typical word error rates within trained domains of less than 5%.
- SensoryCloud Wake Word Revalidation – Leveraging the experience of Sensory's expertise in wake word detection, the cloud-based verification of custom, branded wake words enable up to a 90% reduction in false alarm events.
- SensoryCloud Sound Identification – Offers an extensive library of sounds with a multi-stage approach optimised for speed, efficiency, and accuracy. Developers can quickly train and learn custom sounds, in addition to the standard sounds like alarms, sirens, breaking glass, crying babies, coughs, sneezes, doorbells, and more.
Sensory’s technologies have shipped in over three billion products from hundreds of leading consumer electronics manufacturers including ATT, Hasbro, Huawei, Google, Amazon, Samsung, LG, Mattel, Motorola, Plantronics, GoPro, Sony, Tencent, Garmin, LG, Microsoft, Lenovo, and more. The company also has over 60 issued patents covering speech recognition in consumer electronics, biometric authentication, sensor/speech combinations, wake word technology, and more.