NASA Implements Deepgram for Space Audio Transcription

This case study details NASA's implementation of Deepgram's speech-to-text technology to enhance audio transcription capabilities for various space-related applications. The document outlines the challenges faced by NASA, particularly the difficulty in transcribing noisy audio from space communications. Previous attempts with other speech-to-text providers did not meet the required accuracy threshold until Deepgram's API was utilized. The solution involved creating tailored models capable of handling specific audio types, resulting in word recognition rates of 86-87%. The case study describes four distinct use cases for Deepgram's technology, including space-to-ground communications, training at the Neutral Buoyancy Lab, the Medical Interactive Response Intelligent System (IRIS) chatbot, and an upcoming robotic processing automation system. Each use case highlights the benefits of improved accuracy and efficiency in mission-critical operations, showcasing how Deepgram's technology supports NASA's objectives in space exploration.