https://www.live-feeds.com/feed/ibm-nasa-launch-ai-model-to-help-map-ice-craters-on-moonContinuously updated, source-cited coverage. https://www.live-feeds.com/feed/ibm-nasa-launch-ai-model-to-help-map-ice-craters-on-moonhttps://www.live-feeds.com/feed/ibm-nasa-launch-ai-model-to-help-map-ice-craters-on-moon#u70565Wed, 16 Sep 2026 14:01:39 +0000NASA and IBM have launched the Lunar Foundation Model, a free open-source AI tool designed to map volcanic sites, craters, and ice on the Moon. The model identifies water ice and analyzes data from five Moon-orbiting spacecraft, utilizing a training set of over two million images. This tool is hosted on Hugging Face with its codebase available on GitHub to help astronauts locate critical resources. While some reports indicate the AI improves water ice detection accuracy by 22%, other data suggests it outperforms existing mapping methods by 23%.Why it mattersAccurate mapping of lunar ice is ess https://www.live-feeds.com/feed/ibm-nasa-launch-ai-model-to-help-map-ice-craters-on-moonhttps://www.live-feeds.com/feed/ibm-nasa-launch-ai-model-to-help-map-ice-craters-on-moon#u66662Sun, 13 Sep 2026 02:25:20 +0000IBM stock rose 2% following the release of the NASA-IBM Lunar Foundation Model, an open-source AI designed to map lunar ice, craters, and volcanic features. The model analyzes 17 years of data from the Lunar Reconnaissance Orbiter to identify subsurface ice in polar craters. While previous reports cited a 22% error reduction, newer data indicates the AI beats existing mapping methods by 23%. The codebase is available on GitHub and the model is hosted on Hugging Face to assist astronauts in locating critical resources.Why it mattersMapping water ice is essential for sustainable human presence o https://www.live-feeds.com/feed/ibm-nasa-launch-ai-model-to-help-map-ice-craters-on-moonhttps://www.live-feeds.com/feed/ibm-nasa-launch-ai-model-to-help-map-ice-craters-on-moon#u65526Fri, 11 Sep 2026 22:00:36 +0000NASA and IBM released the NASA-IBM Lunar Foundation Model on September 10, 2026, to help researchers map lunar ice, craters, and volcanic features. The open-source model is hosted on Hugging Face with its codebase available on GitHub. Trained on 17 years of data primarily from the Lunar Reconnaissance Orbiter, the AI analyzes petabytes of information to identify subsurface ice reserves in permanently shadowed polar craters. The system predicts ice locations with 22% less error than the best existing dedicated models.Why it mattersNASA intends to return humans to the Moon in 2028. Mapping water