IBM (NYSE: IBM) inches down in pre trading session on Thursday as the firm announced with NASA’s Marshall Space Flight Center a cooperation to identify new insights in NASA’s huge repository of Earth and geospatial scientific data using IBM’s artificial intelligence (AI) technologies. For the first time, the collaborative endeavor will apply AI foundation model technology to NASA’s Earth-observing satellite data.
Foundation models are AI models that are trained on a large quantity of unlabeled data, may be utilized for a variety of tasks, and can apply knowledge from one context to another. Over the last five years, these models have dramatically improved the field of natural language processing (NLP) technology, and IBM is pioneering uses of foundation models outside language.
Earth observations are being collected at unprecedented rates and volumes, allowing scientists to study and monitor our world. To extract insight from these large data resources, new and novel ways are necessary. The purpose of this effort is to make it easier for academics to examine and draw conclusions from enormous datasets. IBM’s foundation model technology has the ability to accelerate the finding and processing of this data, allowing scientists to expand their understanding of the Earth and respond to climate-related challenges more swiftly.
IBM and NASA want to create a number of new technologies to harvest information from Earth observations. One project will employ NASA’s Harmonized Landsat Sentinel-2 (HLS) dataset to train an IBM geospatial intelligence foundation model, which is a record of land cover and land use changes acquired by Earth-orbiting satellites. This foundation model technology will assist researchers in providing critical analyses of our planet’s environmental systems by analyzing petabytes of satellite data to discover variations in the global footprint of phenomena such as natural disasters, cyclical agricultural yields, and animal habitats.
“The beauty of foundation models is that they have the potential to be utilized for a wide range of downstream applications,” said Rahul Ramachandran, senior research scientist at NASA’s Marshall Space Flight Center in Huntsville, Alabama. “These basic models cannot be built by tiny teams,” he adds. “You need cross-organizational teams to offer diverse viewpoints, resources, and skill sets.”
“Foundation models have proved effective in natural language processing, and it’s time to extend that to other domains and modalities crucial to business and society,” Raghu Ganti, principal researcher at IBM, stated. “Applying foundation models to geographical, event-sequence, time-series, and other non-language variables in Earth science data might provide extremely useful insights and information to a far larger range of academics, corporations, and individuals. Finally, it may enable a greater number of individuals to work on some of our most important climate concerns.”
