Feed aggregator
Risk of bird flu outbreak in cows causing pandemic is less than feared
Boeing Starliner astronauts arrive at launch site for 1st flight test on June 1 (photos)
Astronomy Generates Mountains of Data. That’s Perfect for AI
Consumer-grade AI is finding its way into people’s daily lives with its ability to generate text and images and automate tasks. But astronomers need much more powerful, specialized AI. The vast amounts of observational data generated by modern telescopes and observatories defies astronomers’ efforts to extract all of its meaning.
A team of scientists is developing a new AI for astronomical data called AstroPT. They’ve presented it in a new paper titled “AstroPT: Scaling Large Observation Models for Astronomy.” The paper is available at arxiv.org, and the lead author is Michael J. Smith, a data scientist and astronomer from Aspia Space.
Astronomers are facing a growing deluge of data, which will expand enormously when the Vera Rubin Observatory (VRO) comes online in 2025. The VRO has the world’s largest camera, and each of its images could fill 1500 large-screen TVs. During its ten-year mission, the VRO will generate about 0.5 exabytes of data, which is about 50,000 times more data than is contained in the USA’s Library of Congress.
The VRO’s need for multiple sites to handle all of its data is a testament to the enormous volume of data it will generate. Without effective AI, that data will be stuck in a bottleneck. Image Credit: NOIRLab.Other telescopes with enormous mirrors are also approaching first light. The Giant Magellan Telescope, the Thirty Meter Telescope, and the European Extremely Large Telescope combined will generate an overwhelming amount of data.
Having data that can’t be processed is the same as not having the data at all. It’s basically inert and has no meaning until it’s processed somehow. “When you have too much data, and you don’t have the technology to process it, it’s like having no data,” said Cecilia Garraffo, a computational astrophysicist at the Harvard-Smithsonian Center for Astrophysics.
This is where AstroPT comes in.
AstroPT stands for Astro Pretrained Transformer, where a transformer is a particular type of AI. Transformers can change or transform an input sequence into an output sequence. AI needs to be trained, and AstroPT has been trained on 8.6 million 512 x 512-pixel images from the DESI Legacy Survey Data Release 8. DESI is the Dark Energy Spectroscopic Instrument. DESI studies the effect of Dark Energy by capturing the optical spectra from tens of millions of galaxies and quasars.
AstroPT and similar AI deal with ‘tokens.’ Tokens are visual elements in a larger image that contain meaning. By breaking images down into tokens, an AI can understand the larger meaning of an image. AstroPT can transform individual tokens into coherent output.
AstroPT has been trained on visual tokens. The idea is to teach the AI to predict the next token. The more thoroughly it’s been trained to do that, the better it will perform.
“We demonstrated that simple generative autoregressive models can learn scientifically useful information when pre-trained on the surrogate task of predicting the next 16 × 16 pixel patch in a sequence of galaxy image patches,” the authors write. In this scheme, each image patch is a token.
This image illustrates how the authors trained AstroPT to predict the next token in a ‘spiralised’ sequence of galaxy image patches. It shows the token feed order. “As the galaxies are in the centre of each postage stamp, this set up allows us to seamlessly pretrain and run inference on differently sized galaxy postage stamps,” the authors explain. Image Credit: Smith et al. 2024.One of the obstacles to training AI like AstroPT concerns what AI scientists call the ‘token crisis.’ To be effective, AI needs to be trained on a large number of quality tokens. In a 2023 paper, a separate team of researchers explained that a lack of tokens can limit the effectiveness of some AI, such as LLMs or Large Language Models. “State-of-the-art LLMs require vast amounts of internet-scale text data for pre-training,” the wrote. “Unfortunately, … the growth rate of high-quality text data on the internet is much
slower than the growth rate of data required by LLMs.”
AstroPT faces the same problem: a dearth of quality tokens to train on. Like other AI, it uses LOMs or Large Observation Models. The team says their results so far suggest that AstroPT can solve the token crisis by using data from observations. “This is a promising result that suggests that data taken from the observational sciences would complement data from other domains when used to pre-train a single multimodal LOM, and so points towards the use of observational data as one solution to the ‘token crisis’.”
AI developers are eager to find solutions to the token crisis and other AI challenges.
Without better AI, a data processing bottleneck will prevent astronomers and astrophysicists from making discoveries from the vast quantities of data that will soon arrive. Can AstroPT help?
The authors are hoping that it can, but it needs much more development. They say they’re open to collaborating with others to strengthen AstroPT. To aid that, they followed “current leading community models” as closely as possible. They call it an “open to all project.”
“We took these decisions in the belief that collaborative community development paves the fastest route towards realising an open source web-scale large observation model,” they write.
“We warmly invite potential collaborators to join us,” they conclude.
It’ll be interesting to see how AI developers will keep up with the vast amount of astronomical data coming our way.
The post Astronomy Generates Mountains of Data. That’s Perfect for AI appeared first on Universe Today.
Real-life 'Star Trek' planet was actually just an illusion caused by a 'jittery' star
Parkinson’s disease could be prevented by a recent tetanus vaccine
Parkinson’s disease could be prevented by a recent tetanus vaccine
Quantum time travel: The experiment to 'send a particle into the past'
Quantum time travel: The experiment to 'send a particle into the past'
NASA's James Webb Space Telescope mission — Live updates
Summary of the Fifty-Second U.S.–Japan ASTER Science Team Meeting
8 min read
Summary of the Fifty-Second U.S.–Japan ASTER Science Team MeetingMichael Abrams, NASA/Jet Propulsion Laboratory/California Institute of Technology, mjabrams@jpl.nasa.gov
Yasushi Yamaguchi, Nagoya University/Japan Science and Technology Agency, yasushi@nagoya-u.jp
Introduction
The Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) Science Team (ST) organized a three-day workshop that took place September 11–13, 2023, at the offices of Japan Space Systems (JSS) in Tokyo. Over 40 people from Japan and the U.S. participated in the in-person meeting—some of whom are shown in the Photo below. U.S. participants included members from NASA/Jet Propulsion Laboratory (JPL), NASA’s Land Processes Distributed Active Archive Center (LPDAAC), NASA’s Goddard Space Flight Center (GSFC), University of Arizona (UA), Grace Consulting (GC), and University of Pittsburgh (Pitt). Japanese members included representatives from JSS, Ibaraki University (IU), Nagoya University (NU), University of Tokyo (UT), Geologic Survey of Japan (GSJ), National Institute of Advanced Industrial Science and Technology (AIST), University of Tsukuba (UTs), and Remote Sensing Technology Center of Japan (RESTEC).
The meeting objectives focused on discussing impacts of the 50% budget reductions to the Terra mission (including ASTER) that have been proposed in the NASA Budget for Fiscal Years (FY) 2024–26; revised spacecraft management protocols by the Flight Operations Team; data acquisition status; data calibration and validation; data distribution; status of Level-1 processing interruption; applications; and end-of-mission plans. After summarizing the opening plenary presentations, the remainder of this article provides highlights from meetings of the various ASTER working groups and the closing plenary session.
Photo. Some of the attendees at the fifty-second ASTER STM. Photo credit: Mako Komoda, JSSOpening Plenary Session
Yasushi Yamaguchi [NU] and Michael Abrams [JPL—ASTER ST Leaders from Japan and the U.S., respectively] welcomed participants and reviewed the agenda for the opening plenary and the schedule for the week’s working groups.
Akira Tsuneto [AIST—Vice President], whose office is responsible for the ASTER project, presented a special welcome. As the former Director of Space Industry Office in the Japan Ministry of Economy, Trade and Industry (METI), he was responsible for making ASTER data free to all users.
Michael Abrams [JPL] presented Jason Hendrickson’s [GSFC] slides on the operations status of NASA’s Terra platform—which has changed significantly since the last meeting. The Earth Science Mission Operations (ESMO) Flight Operations Team began implementing “Lights Out Operation,” reducing staff from 24/7 coverage and eliminating the night shift. These changes resulted in a small increase in data gaps and delayed anomaly response. In early 2023 Terra lost two of its 24 solar array shunts. Full power capability remains—however, there is only one spare shunt remaining. Those issues notwithstanding, Terra remains healthy after more than 23 years of operation.
Chris Torbert [LPDAAC] presented ASTER product distribution statistics. The ASTER Global Digital Elevation Model (DEM) continues to be the most ordered product. Torbert discussed the ASTER Preservation Content Specification for the end-of-mission archiving. There is a NASA document that describes the desired content of this archive. As described by the ST at the last meeting, most ASTER data products will be created as real files and placed in a searchable and orderable archive, accessed through NASA’s Earthdata tool, where mission preservation documents for other instruments (e.g., HIRDLS, ICESat/GLAS, TOMS) can be found.
Michael Abrams [JPL] presented highlights of science results based on ASTER data—including the 2023 Earth Science Senior Review. Terra presented its report to NASA Headquarters, but as of this meeting, the response is still pending. However, as stated earlier, a three-year budget reduction of 50% is anticipated.
Hitomi Inada [JSS] presented the status of the ASTER instrument. Although many of the monitored components [e.g., visible-near-infrared (VNIR) pointing motor] have exceeded their original useful life in orbit, they show no signs of decreases in performance. All temperature and current telemetry trends remain straight lines.
Tetsushi Tachikawa [JSS] summarized the status of ASTER observations since the beginning of the mission. He reported that all of the global observation programs are functioning normally, acquiring data as planned. The change of the orbit repeat after the October 2022 constellation exit maneuver has been accommodated in the ASTER scheduler.
Simon Hook [JPL] described the status of the multispectral thermal infrared (TIR) instrument on the ECOsystem Spaceborne Thermal Radiometer Experiment on Space Station (ECOSTRESS) as well as NASA’s future Surface Biology and Geology (SBG) mission, which is part of the planned Earth System Observatory.
Applications Working Group
The applications session offered a sample of the variety of applications that make use of data from ASTER, see examples below. Miyuki Muto [IU] shared her work to estimate the volume of waste in 19 landfills in 11 countries through analysis of ASTER DEM data over the past 20 years. Analysis of data from a site in India showed that the volume of waste increased four-fold over 20 years—see Figure 1. All the other monitored sites showed similar large increases in waste volume.
Figure 1. Google Earth Image of landfill in India [top] and temporal changes in volume from 2001 to 2021 [bottom]. Figure credit: Miyuki Muto and Hideyuki Tonooka, IU Figure credit: Miyuki Muto and Hideyuki Tonooka, IUMichael Ramsey [Pitt] discussed detecting volcanic eruption precursors using the entire ASTER TIR archive for six selected volcanoes: Etna, Fuego, Kliuchevskoi, Lascar, Vulcano, and Popocatepetl—four of these are shown in Figure 2. He and his students developed statistical methods to detect both low- and high-temperature anomalies. The team performed a cluster analysis on four volcanoes. By calculating and plotting heat flux versus mean temperature-above-background versus maximum temperature-above-background, clusters for eruption styles can be identified—see Figure 2. These results offer potential applicability to other volcanoes.
Figure 2. Three-dimensional plots show heat flux and temperature plots (further explained in the text) for hundreds of ASTER TIR scenes for four volcanoes, revealing differences related to eruptive styles. The lower cluster (blue) indicated fumarole and passive degassing; the medium cluster (red) correlated with domes and explosive and small lava flows; and the high clusters (green) correlated with large lava flows. Figure credit: Michael Ramsey/PittCalibration/Validation Working Group
This working group monitors the radiometric performance of ASTER’s VNIR and TIR instruments. The team performs calibration and validation of these instruments by analysis of onboard calibration lamps or blackbody, as well as measurements of pseudo-invariant ground targets during field campaigns. No changes in instrument performance were found based on validation activities during the past year. The radiometric calibration coefficients will remain unchanged for the foreseeable future.
Temperature–Emissivity Working Group
The Temperature–Emissivity Working Group focuses on ASTER’s kinetic temperature and emissivity (T–E) products and their applications, including monitoring instrument performance and calibration. They also review the status of the nighttime TIR global map program. In situ measurement campaigns in Japan and the U.S. use lakes and dry lake beds for ground-based calibration campaigns. Recent campaign results indicate that the TIR instrument perform within required calibration limits—see Figure 3. The team also noted the successful completion of the Visible Infrared Imaging Radiometer Suite (VIIRS)–ASTER 375-m (~1230-ft) near-real-time land-surface temperature algorithm using ASTER emissivity for corrections. Review of the thermal global mapping acquisition program indicated that it was proceeding as planned with no changes needed.
Figure 3. ASTER and Landsat 8 and 9 data provide a way to compare the satellite-derived temperature and lake surface measured temperature. ASTER mean difference for all five bands is less than 0.5 °C (~0.9 °F). On the Y axis, BT stands for Brightness Temperature. Figure credit: Remote Sensing Technology Center of Japan/Soushi Kato Figure credit: Remote Sensing Technology Center of Japan/Soushi KatoOperations and Mission Planning Working Group
The Operations and Mission Planning working group oversees and reviews the acquisition programs executed by the ASTER scheduler. The working group schedules ASTER data acquisitions daily to accommodate ASTER’s average 8% duty cycle. An automated program selects 600–700 daily scenes from the more than 3000 in the request archive.
Tetsushi Tachikawa [JSS] reviewed the status of acquisition scheduling. Urgent observations receive the highest priority and can be scheduled close to acquisition time. Approximately 70 scenes are programmed per month—with over 95% acquisition success. By contrast, global mapping data acquisitions receive the lowest priority and fill in the scenes for the daily quota. The objective is for ASTER to acquire at least one cloud-free image for every place on Earth. Due to persistent cloud cover, success is typically ~85%. The group restarts the program after several years, with the next scheduled restart in October 2024. The thermal group submits aerial requirements to acquire global nighttime coverage with the thermal bands, which will continue as scheduled. There are also acquisition programs that focus on islands, volcanoes, glaciers, and cloudy areas. The global volcano image acquisition program will continue with no change to the observation parameters. Acquisition of images of islands and over cloudy areas will also continue in current form. The global glacier acquisition program will be modified to change the VNIR gain settings to optimize images over snow and ice.
Chris Torbert [LPDAAC] reported that software fixes were ongoing for the (currently non-functional) expedited data processing at the LPDAAC.
Closing Plenary Session
Each working group chairperson summarized the presentations, discussions, and recommendations that occurred during each session. Consensus holds the ASTER instrument is operating normally, with no indications of any component failures. The backlog of unprocessed scenes resulting from the 2022 constellation exit maneuver impact on production software should clear by early October 2023. The closing highlighted the impact of the 50% budget reduction on the Flight Operation Team at GSFC with only a small increase in lost data (1–2%) due to the absence of operators to attempt immediate recovery.
Conclusion
The fifty-second ASTER ST Meeting successfully covered all of the critical issues introduced during the opening plenary session. Working groups updated instrument scheduling, instrument performance, archiving plans, and new applications. The plan is for the 2024 meeting to take place at the same venue in Tokyo.
NASA Lucy Images Reveal Asteroid Dinkinesh to be Surprisingly Complex
4 min read
Preparations for Next Moonwalk Simulations Underway (and Underwater)Images from the November 2023 flyby of asteroid Dinkinesh by NASA’s Lucy spacecraft show a trough on Dinkinesh where a large piece — about a quarter of the asteroid — suddenly shifted, a ridge, and a separate contact binary satellite (now known as Selam). Scientists say this complicated structure shows that Dinkinesh and Selam have significant internal strength and a complex, dynamic history.
Panels a, b, and c each show stereographic image pairs of the asteroid Dinkinesh taken by the NASA Lucy Spacecraft’s L’LORRI Instrument in the minutes around closest approach on Nov. 1, 2023. The yellow and rose dots indicate the trough and ridge features, respectively. These images have been sharpened and processed to enhance contrast. Panel d shows a side view of Dinkinesh and its satellite Selam taken a few minutes after closest approach.NASA/GSFC/SwRI/Johns Hopkins APL/NOIRLab“We want to understand the strengths of small bodies in our solar system because that’s critical for understanding how planets like Earth got here,” said Hal Levison, Lucy principal investigator at the Boulder, Colorado, branch of the Southwest Research Institute in San Antonio, Texas. “Basically, the planets formed when zillions of smaller objects orbiting the Sun, like asteroids, ran into each other. How objects behave when they hit each other, whether they break apart or stick together, has a lot to do with their strength and internal structure.” Levison is lead author of a paper on these observations published May 29 in Nature.
On November 1, 2023, NASA’s Lucy spacecraft flew by the main-belt asteroid Dinkinesh. Now, the mission has released pictures from Lucy’s Long Range Reconnaissance Imager taken over a roughly three-hour period, providing the best views of the asteroid to date. During the flyby, Lucy discovered that Dinkinesh has a small moon, which the mission named “Selam,” a greeting in the Amharic language meaning “peace.” Lucy is the first mission designed to visit the Jupiter Trojans, two swarms of asteroids trapped in Jupiter’s orbit that may be “fossils” from the era of planet formation. Credit: NASA’s Goddard Space Flight Center. Download this video and more at: https://svs.gsfc.nasa.gov/14596/Researchers think that Dinkinesh is revealing its internal structure by how it has responded to stress. Over millions of years rotating in the sunlight, the tiny forces coming from the thermal radiation emitted from the asteroid’s warm surface generated a small torque that caused Dinkinesh to gradually rotate faster, building up centrifugal stresses until part of the asteroid shifted into a more elongated shape. This event likely caused debris to enter into a close orbit, which became the raw material that produced the ridge and satellite.
Stereo movie of asteroid Dinkinesh from NASA’s Lucy spacecraft flyby on Nov. 1, 2023.NASA/GSFC/SwRI/Johns Hopkins APL/NOIRLab/Brian May/Claudia ManzoniIf Dinkinesh were much weaker, more like a fluid pile of sand, its particles would have gradually moved toward the equator and flown off into orbit as it spun faster. However, the images suggest that it was able to hold together longer, more like a rock, with more strength than a fluid, eventually giving way under stress and fragmenting into large pieces. (Although the amount of strength needed to fragment a small asteroid like Dinkinesh is miniscule compared to most rocks on Earth.)
“The trough suggests an abrupt failure, more an earthquake with a gradual buildup of stress and then a sudden release, instead of a slow process like a sand dune forming,” said Keith Noll of NASA’s Goddard Space Flight Center in Greenbelt, Maryland, project scientist for Lucy and a co-author of the paper.
“These features tell us that Dinkinesh has some strength, and they let us do a little historical reconstruction to see how this asteroid evolved,” said Levison. “It broke, things moved apart and formed a disk of material during that failure, some of which rained back onto the surface to make the ridge.”
The researchers think some of the material in the disk formed the moon Selam, which is actually two objects touching each other, a configuration called a contact binary. Details of how this unusual moon formed remain mysterious.
Stereo movie of Selam from NASA’s Lucy spacecraft flyby on Nov. 1, 2023.NASA/GSFC/SwRI/Johns Hopkins APL/NOIRLab/Brian May/Claudia ManzoniDinkinesh and its satellite are the first two of 11 asteroids that Lucy’s team plans to explore over its 12-year journey. After skimming the inner edge of the main asteroid belt, Lucy is now heading back toward Earth for a gravity assist in December 2024. That close flyby will propel the spacecraft back through the main asteroid belt, where it will observe asteroid Donaldjohanson in 2025, and then on to the first of the encounters with the Trojan asteroids that lead and trail Jupiter in its orbit of the Sun beginning in 2027.
Lucy’s principal investigator is based out of the Boulder, Colorado, branch of Southwest Research Institute, headquartered in San Antonio. NASA’s Goddard Space Flight Center in Greenbelt, Maryland, provides overall mission management, systems engineering, and safety and mission assurance. Lockheed Martin Space in Littleton, Colorado, built and operates the spacecraft. Lucy is the 13th mission in NASA’s Discovery Program. NASA’s Marshall Space Flight Center in Huntsville, Alabama, manages the Discovery Program for the Science Mission Directorate at NASA Headquarters in Washington.
For more information about NASA’s Lucy mission, visit:
https://science.nasa.gov/mission/lucy
Share Details Last Updated May 29, 2024 EditorWilliam SteigerwaldContactWilliam Steigerwaldwilliam.a.steigerwald@nasa.govLocationGoddard Space Flight Center Related Terms Explore More 1 min read What are the Trojan Asteroids? We Asked a NASA ScientistWhat are the Trojan asteroids? These mysterious space rocks have been gravitationally trapped in Jupiter’s…
Article 2 years ago 3 min read Asteroid Fast Facts Article 10 years ago 13 min read NASA’s Lucy Mission: A Journey to the Young Solar SystemNASA’s Lucy spacecraft launched Oct. 16, 2021, on a 12-year journey to Jupiter’s Trojan asteroids.
Article 3 years ago
