Harnessing the Power of AI to Unlock the Unused Potential of NASA Earth Data

IBM AI helps NASA researchers mine Earth data

NASA receives a huge amount of data every day. Some data is processed instantly, but much of it is stored to be processed years down the road. Researchers must change this situation if they are to use the data for investigating critical issues that have dynamically changing characteristics, such as global climate change. NASA’s Marshall Space Flight Center has announced a joint program with IBM Research that will process NASA data using IBM Research’s AI foundation technology.

This task can be put into perspective by comparing the GPT-3 dataset, which was used to develop the ChatGPT AI platform, and recently passed the Wharton MBA examination, with the 45TB (terabytes), of data. NASA estimates its data set to be up to 250PB (petabytes). NASA’s data set is 5,000 times bigger than GPT-3, with 1PB equaling 1,000TB.

IBM had estimated that 90% (or more) of the data collected was never used. In their invitation to the press, IBM and NASA stated that \”currently, half of the scientific findings are derived from archived data. This makes it difficult for researchers and scientists to study constantly evolving threats like climate change.\” To efficiently mine the vast amount of archival information, AI is needed. IBM Research’s cloud resources, collective experience of its AI experts and AI foundation model will enable NASA to filter and analyze Earth-Science data in days, not years.