First generative AI model deployed on ISS: Know its significance
Booz Allen Hamilton, a supplier of AI solutions to the US government, has achieved a milestone by deploying a generative AI large language model (LLM) on a supercomputer aboard the International Space Station (ISS). The company used Hewlett Packard Enterprise's (HPE) Spaceborne Computer-2 for this deployment. The project was executed over an eight-week period and is set to revolutionize remote data ingestion as well as retrieval-augmented generation.
AI model to enhance data retrieval
The newly deployed AI LLM is expected to significantly improve the efficiency of data retrieval for edge-deployed personnel. It will aid them in accurately interpreting information and resolving complex issues using natural language processing in space. This innovative application of AI technology marks a significant advancement in space-based data management and interpretation.
Booz Allen's commitment to AI in space
Chris Bogdan, Executive VP at Booz Allen and leader of the firm's space business, expressed his enthusiasm about the successful deployment. He stated that this capability "unlocks the potential for on-orbit generative AI to integrate and develop mission-critical solutions." Bogdan emphasized Booz Allen's commitment to pushing the boundaries of what is possible with AI, and other mission-critical technologies in space.
AI deployment potential in extreme conditions
The successful deployment of the AI model on the ISS could pave the way for its use in other challenging environments, both on Earth and in space. The HPE Spaceborne Computer-2, installed earlier this year on the ISS, is designed to handle data center-level processing and high-performance computing (HPC). This includes workloads related to AI and machine learning (ML).
Supercomputer revolutionizes data processing in space
The HPE Spaceborne Computer-2 has been upgraded with over 130TB of flash-based storage from KIOXIA, marking the most storage ever sent to the space station on a single mission. Unlike traditional methods where data collected in space was transmitted to Earth for processing, this supercomputer allows data evaluation in low Earth orbit in near-real time. This results in a significant reduction in download size and times, according to HPE.
