Members Can Post Anonymously On This Site
-
Similar Topics
-
By NASA
A collaboration between IMPACT and IBM has produced INDUS, a comprehensive suite of large language models (LLMs) tailored for the domains of Earth science, biological and physical sciences, heliophysics, planetary sciences, and astrophysics and trained using curated scientific corpora drawn from diverse data sources. Kaylin Bugbee (ST11), team lead of NASA’s Science Discovery Engine (SDE), spoke to the benefit INDUS offers to existing applications: “Large language models are rapidly changing the search experience. The Science Discovery Engine, a unified, insightful search interface for all of NASA’s open science data and information, has prototyped integrating INDUS into its search engine. Initial results have shown that INDUS improved the accuracy and relevancy of the returned results.”
The INDUS models are openly available on Hugging Face. For the benefit of the scientific community, the team has released the developed models and will release the benchmark datasets that span named entity recognition for climate change, extractive QA for Earth science, and information retrieval for multiple domains. A paper on INDUS, “INDUS: Effective and Efficient Language Models for Scientific Applications,” is available at https://arxiv.org/pdf/2405.10725.
View the full article
-
By NASA
SatSummit brings together leaders in the satellite industry and global development experts for two days of presentations and discussions on using satellite data to address critical development challenges. Rahul Ramachandran (ST11/IMPACT) participated in a panel focused on large earth foundation models, offering an overview of AI foundation models and their potential for societal good. He detailed NASA’s approach to building these models and the agency’s overall strategy, underscoring their importance in advancing Earth science and global development initiatives.
View the full article
-
By NASA
4 min read
Preparations for Next Moonwalk Simulations Underway (and Underwater)
This image shows an aviation version of a smartphone navigation app that makes suggestions for an aircraft to fly an alternate, more efficient route. The new trajectories are based on information available from NASA’s Digital Information Platform and processed by the Collaborative Departure Digital Rerouting tool.NASA Just like your smartphone navigation app can instantly analyze information from many sources to suggest the best route to follow, a NASA-developed resource is now making data available to help the aviation industry do the same thing.
To assist air traffic managers in keeping airplanes moving efficiently through the skies, information about weather, potential delays, and more is being gathered and processed to support decision making tools for a variety of aviation applications.
Appropriately named the Digital Information Platform (DIP), this living database hosts key data gathered by flight participants such as airlines or drone operators. It will help power additional tools that, among other benefits, can save you travel time.
Ultimately, the aviation industry… and even the flying public, will benefit from what we develop.
Swati Saxena
NASA Aerospace Engineer
“Through DIP we’re also demonstrating how to deliver digital services for aviation users via a modern cloud-based, service-oriented architecture,” said Swati Saxena, DIP project manager at NASA’s Ames Research Center in California.
The intent is not to compete with others. Instead, the hope is that industry will see DIP as a reference they can use in developing and implementing their own platforms and digital services.
“Ultimately, the aviation industry – the Federal Aviation Administration, commercial airlines, flight operators, and even the flying public – will benefit from what we develop,” Saxena said.
The platform and digital services have even more benefits than just saving some time on a journey.
For example, NASA recently collaborated with airlines to demonstrate a traffic management tool that improved traffic flow at select airports, saving thousands of pounds of jet fuel and significantly reducing carbon emissions.
Now, much of the data gathered in collaboration with airlines and integrated on the platform is publicly available. Users who qualify can create a guest account and access DIP data at a new website created by the project.
It’s all part of NASA’s vision for 21st century aviation involving revolutionary next-generation future airspace and safety tools.
Managing Future Air Traffic
During the 2030s and beyond, the skies above the United States are expected to become much busier.
Facing this rising demand, the current National Airspace System – the network of U.S. aviation infrastructure including airports, air navigation facilities, and communications – will be challenged to keep up. DIP represents a key piece of solving that challenge.
NASA’s vision for future airspace and safety involves new technology to create a highly automated, safe, and scalable environment.
What this vision looks like is a flight environment where many types of vehicles and their pilots, as well as air traffic managers, use state-of-the-art automated tools and systems that provide highly detailed and curated information.
These tools leverage new capabilities like machine learning and artificial intelligence to streamline efficiency and handle the increase in traffic expected in the coming decades.
Digital Services Ecosystem in Action
To begin implementing this new vision, our aeronautical innovators are evaluating their platform, DIP, and services at several airports in Texas. This initial stage is a building block for larger such demonstrations in the future.
“These digital services are being used in the live operational environment by our airline partners to improve efficiency of the current airspace operations,” Saxena said. “The tools are currently in use in the Dallas/Fort Worth area and will be deployed in the Houston airspace in 2025.”
The results from these digital tools are already making a difference.
Proven Air Traffic Results
During 2022, a NASA machine learning-based tool named Collaborative Digital Departure Rerouting, designed to improve the flow of air traffic and prevent flight delays, saved more than 24,000 lbs. (10,886 kg.) of fuel by streamlining air traffic in the Dallas area.
If such tools were used across the entire country, the improvements made in efficiency, safety, and sustainability would make a notable difference to the flying public and industry.
“Continued agreements with airlines and the aviation industry led to the creation and expansion of this partnership ecosystem,” Saxena said. “There have been benefits across the board.”
DIP was developed under NASA’s Airspace Operations and Safety Program.
Learn about NASA’s Collaborative Digital Departure Rerouting tool and how it uses information from the Digital Information Platform to provide airlines with routing options similar to how drivers navigate using cellphone apps. About the Author
John Gould
Aeronautics Research Mission DirectorateJohn Gould is a member of NASA Aeronautics' Strategic Communications team at NASA Headquarters in Washington, DC. He is dedicated to public service and NASA’s leading role in scientific exploration. Prior to working for NASA Aeronautics, he was a spaceflight historian and writer, having a lifelong passion for space and aviation.
Facebook logo @NASA@NASAaero@NASA_es @NASA@NASAaero@NASA_es Instagram logo @NASA@NASAaero@NASA_es Linkedin logo @NASA Explore More
2 min read NASA Prepares for Air Taxi Passenger Comfort Studies
Article 2 weeks ago 2 min read Hypersonic Technology Project Overview
Article 3 weeks ago 2 min read Hypersonics Technical Challenges
Article 3 weeks ago Keep Exploring Discover More Topics From NASA
Missions
Artemis
Aeronautics STEM
Explore NASA’s History
Share
Details
Last Updated Jul 12, 2024 EditorJim BankeContactJim Bankejim.banke@nasa.gov Related Terms
Aeronautics Aeronautics Research Mission Directorate Air Traffic Management – Exploration Air Traffic Solutions Airspace Operations and Safety Program View the full article
-
By NASA
The Goldstone Solar System Radar, part of NASA’s Deep Space Network, made these observations of the recently discovered 500-foot-wide (150-meter-wide) asteroid 2024 MK, which made its closest approach — within about 184,000 miles (295,000 kilometers) of Earth — on June 29.NASA/JPL-Caltech The Deep Space Network’s Goldstone planetary radar had a busy few days observing asteroids 2024 MK and 2011 UL21 as they safely passed Earth.
Scientists at NASA’s Jet Propulsion Laboratory in Southern California recently tracked two asteroids as they flew by our planet. One turned out to have a little moon orbiting it, while the other had been discovered only 13 days before its closest approach to Earth. There was no risk of either near-Earth object impacting our planet, but the radar observations taken during these two close approaches will provide valuable practice for planetary defense, as well as information about their sizes, orbits, rotation, surface details, and clues as to their composition and formation.
Passing Earth on June 27 at a distance of 4.1 million miles (6.6 million kilometers), or about 17 times the distance between the Moon and Earth, the asteroid 2011 UL21 was discovered in 2011 by the NASA-funded Catalina Sky Survey, in Tucson, Arizona. But this is the first time it has come close enough to Earth to be imaged by radar. While the nearly mile-wide (1.5-kilometer-wide) object is classified as being potentially hazardous, calculations of its future orbits show that it won’t pose a threat to our planet for the foreseeable future.
Because close approaches by asteroids the size of 2024 MK are relatively rare, JPL’s planetary radar team gathered as much information about the near-Earth object as possible. This mosaic shows the spinning asteroid in one-minute increments about 16 hours after its closest approach with Earth.NASA/JPL-Caltech Using the Deep Space Network’s 230-foot-wide (70-meter) Goldstone Solar System Radar, called Deep Space Station 14 (DSS-14), near Barstow, California, JPL scientists transmitted radio waves to the asteroid and received the reflected signals by the same antenna. In addition to determining the asteroid is roughly spherical, they discovered that it’s a binary system: A smaller asteroid, or moonlet, orbits it from a distance of about 1.9 miles (3 kilometers).
“It is thought that about two-thirds of asteroids of this size are binary systems, and their discovery is particularly important because we can use measurements of their relative positions to estimate their mutual orbits, masses, and densities, which provide key information about how they may have formed,” said Lance Benner, principal scientist at JPL who helped lead the observations.
These seven radar observations by the Deep Space Network’s Goldstone Solar System Radar shows the mile-wide asteroid 2011 UL21 during its June 27 close approach with Earth from about 4 million miles away. The asteroid and its small moon (a bright dot at the bottom of the image) are circled in white.NASA/JPL-Caltech Second Close Approach
Two days later, on June 29, the same team observed the asteroid 2024 MK pass our planet from a distance of only 184,000 miles (295,000 kilometers), or slightly more than three-quarters of the distance between the Moon and Earth. About 500 feet (150 meters) wide, this asteroid appears to be elongated and angular, with prominent flat and rounded regions. For these observations, the scientists also used DSS-14 to transmit radio waves to the object, but they used Goldstone’s 114-foot (34-meter) DSS-13 antenna to receive the signal that bounced off the asteroid and came back to Earth. The result of this “bistatic” radar observation is a detailed image of the asteroid’s surface, revealing concavities, ridges, and boulders about 30 feet (10 meters) wide.
Close approaches of near-Earth objects the size of 2024 MK are relatively rare, occurring about every couple of decades, on average, so the JPL team sought to gather as much data about the object as possible. “This was an extraordinary opportunity to investigate the physical properties and obtain detailed images of a near-Earth asteroid,” said Benner.
This sunset photo shows NASA’s Deep Space Station 14 (DSS-14), the 230-foot-wide (70-meter) antenna at the Goldstone Deep Space Communications Complex near Barstow, California.NASA/JPL-Caltech The asteroid 2024 MK was first reported on June 16 by the NASA-funded Asteroid Terrestrial-impact Last Alert System (ATLAS) at Sutherland Observing Station in South Africa. Its orbit was changed by Earth’s gravity as it passed by, reducing its 3.3-year orbital period around the Sun by about 24 days. Although it is classified as a potentially hazardous asteroid, calculations of its future motion show that it does not pose a threat to our planet for the foreseeable future.
The Goldstone Solar System Radar Group is supported by NASA’s Near-Earth Object Observations Program within the Planetary Defense Coordination Office at the agency’s headquarters in Washington. Managed by JPL, the Deep Space Network receives programmatic oversight from Space Communications and Navigation program office within the Space Operations Mission Directorate, also at NASA Headquarters.
More information about planetary radar and near-Earth objects can be found at:
https://www.jpl.nasa.gov/asteroid-watch
News Media Contact
Ian J. O’Neill
Jet Propulsion Laboratory, Pasadena, Calif.
818-354-2649
ian.j.oneill@jpl.nasa.gov
2024-097
Share
Details
Last Updated Jul 03, 2024 Related Terms
Near-Earth Asteroid (NEA) Asteroids Deep Space Network Jet Propulsion Laboratory Planetary Defense Planetary Defense Coordination Office Potentially Hazardous Asteroid (PHA) Space Communications & Navigation Program Explore More
3 min read NASA’s ECOSTRESS Maps Burn Risk Across Phoenix Streets
Article 19 hours ago 5 min read NASA Asteroid Experts Create Hypothetical Impact Scenario for Exercise
Article 23 hours ago 5 min read NASA’s NEOWISE Infrared Heritage Will Live On
Article 2 days ago View the full article
-
By NASA
4 min read
NASA-IBM Collaboration Develops INDUS Large Language Models for Advanced Science Research
Named for the southern sky constellation, INDUS (stylized in all caps) is a comprehensive suite of large language models supporting five science domains. NASA By Derek Koehl
Collaborations with private, non-federal partners through Space Act Agreements are a key component in the work done by NASA’s Interagency Implementation and Advanced Concepts Team (IMPACT). A collaboration with International Business Machines (IBM) has produced INDUS, a comprehensive suite of large language models (LLMs) tailored for the domains of Earth science, biological and physical sciences, heliophysics, planetary sciences, and astrophysics and trained using curated scientific corpora drawn from diverse data sources.
INDUS contains two types of models; encoders and sentence transformers. Encoders convert natural language text into numeric coding that can be processed by the LLM. The INDUS encoders were trained on a corpus of 60 billion tokens encompassing astrophysics, planetary science, Earth science, heliophysics, biological, and physical sciences data. Its custom tokenizer developed by the IMPACT-IBM collaborative team improves on generic tokenizers by recognizing scientific terms like biomarkers and phosphorylated. Over half of the 50,000-word vocabulary contained in INDUS is unique to the specific scientific domains used for its training. The INDUS encoder models were used to fine tune the sentence transformer models on approximately 268 million text pairs, including titles/abstracts and questions/answers.
By providing INDUS with domain-specific vocabulary, the IMPACT-IBM team achieved superior performance over open, non-domain specific LLMs on a benchmark for biomedical tasks, a scientific question-answering benchmark, and Earth science entity recognition tests. By designing for diverse linguistic tasks and retrieval augmented generation, INDUS is able to process researcher questions, retrieve relevant documents, and generate answers to the questions. For latency sensitive applications, the team developed smaller, faster versions of both the encoder and sentence transformer models.
Validation tests demonstrate that INDUS excels in retrieving relevant passages from the science corpora in response to a NASA-curated test set of about 400 questions. IBM researcher Bishwaranjan Bhattacharjee commented on the overall approach: “We achieved superior performance by not only having a custom vocabulary but also a large specialized corpus for training the encoder model and a good training strategy. For the smaller, faster versions, we used neural architecture search to obtain a model architecture and knowledge distillation to train it with supervision of the larger model.”
NASA Chief Scientist Kate Calvin gives remarks in a NASA employee town hall on how the agency is using and developing Artificial Intelligence (AI) tools to advance missions and research, Wednesday, May 22, 2024, at the NASA Headquarters Mary W. Jackson Building in Washington. The INDUS suite of models will help facilitate the agency’s AI goals. NASA/Bill Ingalls INDUS was also evaluated using data from NASA’s Biological and Physical Sciences (BPS) Division. Dr. Sylvain Costes, the NASA BPS project manager for Open Science, discussed the benefits of incorporating INDUS: “Integrating INDUS with the Open Science Data Repository (OSDR) Application Programming Interface (API) enabled us to develop and trial a chatbot that offers more intuitive search capabilities for navigating individual datasets. We are currently exploring ways to improve OSDR’s internal curation data system by leveraging INDUS to enhance our curation team’s productivity and reduce the manual effort required daily.”
At the NASA Goddard Earth Sciences Data and Information Services Center (GES-DISC), the INDUS model was fine-tuned using labeled data from domain experts to categorize publications specifically citing GES-DISC data into applied research areas. According to NASA principal data scientist Dr. Armin Mehrabian, this fine-tuning “significantly improves the identification and retrieval of publications that reference GES-DISC datasets, which aims to improve the user journey in finding their required datasets.” Furthermore, the INDUS encoder models are integrated into the GES-DISC knowledge graph, supporting a variety of other projects, including the dataset recommendation system and GES-DISC GraphRAG.
Kaylin Bugbee, team lead of NASA’s Science Discovery Engine (SDE), spoke to the benefit INDUS offers to existing applications: “Large language models are rapidly changing the search experience. The Science Discovery Engine, a unified, insightful search interface for all of NASA’s open science data and information, has prototyped integrating INDUS into its search engine. Initial results have shown that INDUS improved the accuracy and relevancy of the returned results.”
INDUS enhances scientific research by providing researchers with improved access to vast amounts of specialized knowledge. INDUS can understand complex scientific concepts and reveal new research directions based on existing data. It also enables researchers to extract relevant information from a wide array of sources, improving efficiency. Aligned with NASA and IBM’s commitment to open and transparent artificial intelligence, the INDUS models are openly available on Hugging Face. For the benefit of the scientific community, the team has released the developed models and will release the benchmark datasets that span named entity recognition for climate change, extractive QA for Earth science, and information retrieval for multiple domains. The INDUS encoder models are adaptable for science domain applications, and the INDUS retriever models support information retrieval in RAG applications.
A paper on INDUS, “INDUS: Effective and Efficient Language Models for Scientific Applications,” is available on arxiv.org.
Learn more about the Science Discovery Engine here.
Share
Details
Last Updated Jun 24, 2024 Related Terms
Open Science Explore More
4 min read Marshall Research Scientist Enables Large-Scale Open Science
Article
5 days ago
2 min read NASA’s Repository Supports Research of Commercial Astronaut Health
Article
2 weeks ago
4 min read NASA, IBM Research to Release New AI Model for Weather, Climate
Article
1 month ago
Keep Exploring Discover Related Topics
Missions
Humans in Space
Climate Change
Solar System
View the full article
-
-
Check out these Videos
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.