IBM and Hugging Face Release Climate Science Artificial Intelligence Foundation Model

2023-08-04

In order to democratize the artificial intelligence technology in climate science, IBM and Hugging Face have announced the launch of watsonx.ai geospatial foundation. This geospatial model is built based on NASA's satellite data and will become the largest model on the Hugging Face platform, as well as the first open-source AI foundation model developed in collaboration with NASA. Jeff Boudier, the Product and Growth Lead at Hugging Face, emphasized the importance of information sharing and collaboration in driving advancements in artificial intelligence. The release of open-source AI and model datasets is crucial to ensure that as many people as possible benefit from artificial intelligence. Due to the rapid changes in environmental conditions, climate science faces ongoing challenges and requires access to the latest data. Despite the abundance of data, scientists and researchers still face difficulties in effectively analyzing large datasets. NASA expects the data from new missions to reach 250,000 TB by 2024. To address this issue, IBM signed the Space Act Agreement with NASA earlier this year, aiming to build an AI foundation model for geospatial data. By publicly providing this geospatial foundation model on the Hugging Face platform, both companies aim to promote collaboration and accelerate progress in climate and earth science. Sriram Raghavan, Vice President of IBM Research AI, commented, "The role of open-source technology in accelerating discoveries in critical areas such as climate change has become increasingly clear. By combining IBM's flexible and reusable AI system with NASA's Earth satellite database and making it available on the leading open-source AI platform, Hugging Face, we can leverage the power of collaboration to implement faster and more impactful solutions to improve our planet." The geospatial model trained by IBM and NASA using the Harmonized Landsat Sentinel-2 satellite data (HLS) for one year within the continental United States has shown promising results. With only half of the labeled data, it demonstrated a 15% improvement compared to the current state-of-the-art techniques. With further optimization, this model can be adapted to various tasks such as deforestation tracking, crop yield prediction, and greenhouse gas detection. IBM's collaboration with NASA in building AI models aligns with NASA's decade-long open science initiative, promoting a more accessible and inclusive scientific community. Together with other federal agencies, NASA has designated 2023 as the Year of Open Science, celebrating the benefits of sharing data, information, and knowledge. Kevin Murphy, NASA's Chief Data Scientist, stated, "We believe that foundation models have the potential to change the way we analyze observational data and help us better understand our planet. By open-sourcing such models and making them available worldwide, we hope to expand their impact." The geospatial model leverages IBM's foundational model technology and is part of IBM's broader initiative to create and train transferable AI models to meet the needs of different tasks. In June of this year, IBM launched watsonx, an AI and data platform aimed at extending and accelerating the impact of advanced artificial intelligence through trusted data. The commercial version of the geospatial model will be integrated into IBM watsonx and will be available later this year through IBM's Environmental Intelligence Suite (EIS). By harnessing the power of open-source technology, this latest collaboration aims to effectively address climate challenges and contribute to the sustainable development of the planet.

RECENT AI NEWS

RECENT AI TOOLS