IBM's AI technology is shifting its attention toward Earth science as its latest research goal. NASA and IBM have formed a collaboration that utilizes NASA's geospatial data and historical imagery to gain an understanding of the changes in the Earth and its climate through time. IBM's AI will be processing thousands of terabytes of NASA's Earth science data and imagery and making the process and findings available for free to researchers.
The AI foundation mentioned in a statement that their models are trained on massive amounts of unlabeled data, connecting different pieces of information. The partnership between NASA and IBM aims to streamline NASA's extensive catalog of scientific data, making complex searches simpler and providing more accessible information on the earth's changing climate.
A Climate Change Collaboration
NASA currently holds 70 petabytes of Earth science data, which is expected to increase four-fold with the upcoming launches of the Surface Water and Ocean Topography (SWOT) and NISAR missions by 2024. This is the first instance that IBM applies its AI technology to an Earth science project. The significance of this move is heightened by the growing impact of climate change, which is causing an increase in extreme weather events globally.
The need for prompt, accurate, and comprehensive geospatial data has become imperative. Another outcome from the partnership between NASA and IBM is expected to be a readily searchable collection of Earth science literature. IBM has created an NLP model trained on almost 300,000 journal articles related to Earth science to arrange the literature and make it easier to find new information, as reported by Payload.
This model is one of the largest AI workloads trained on Red Hat's OpenShift software and utilizes IBM's open-source, multilingual question-answering system, PrimeQA. In addition to serving as a resource for researchers, the Earth science language model could also be integrated into NASA's scientific data management and stewardship processes. According to Rahul Ramachandran, a senior research scientist at NASA, the model has the potential to aid disaster responders in promptly mapping major floods or evaluating damage to buildings after severe weather conditions.
IBM AI, NASA's Data
Ramachandran stated that this is the envisioned outcome if the project proves successful. The partnership between IBM and NASA will collaborate to achieve their goals. IBM will allocate resources to construct two AI foundation models utilizing its machine learning technology. NASA will make its data accessible and provide experts in geospatial data and Earth science to assist the model and its developers.
Initially, IBM will begin by analyzing reports using AI technology to uncover a fundamental structure and categorize the information thematically. This is aimed at making it easier for researchers to navigate existing scientific knowledge, according to VentureBeat.
The Harmonized Landsat-Sentinel 2 (HLS2) dataset is a collaboration between NASA and the US Geological Survey to monitor land usage changes over time. For its second effort, IBM will construct an AI foundation model using the HLS2 data, which can then be utilized for various purposes, including disaster prediction, monitoring, and resource management.
RELATED ARTICLE : NASA and IBM Join Forces to Make...Apps
Check out more news and information on Technology in Science Times.
