Space/Planets

From
Jump to: navigation, search

Youtube search... ...Google search ...News search

Google Cloud, NASA FDL & the SETI Institute in search for life on other planets
In search for life on other planets, Google Cloud is partnering up with NASA FDL and SETI Institute. NASA is utilizing Google Cloud to simulate as much data as possible, reading patterns and forming connections from thousands of datasets. Watch the amazing conversation between Massimo Mascaro, Technical Director for Applied AI at Google Cloud and Seth Shostak, Senior Astronomer of the SETI Institute, about Artificial Intelligence, Machine Learning, and all the work behind of partnership of Google Cloud & NASA FDL.

AI and Space Exploration | Intel® AI Interplanetary Show | Intel Software
Bill Nye, Robert Picardo, and Intel’s Hanlin Tang begin our journey by discussing how today’s AI technology helps us explore the solar system.

Jake Vanderplas - Keynote - PyCon 2017
Slides can be found at: https://speakerdeck.com/pycon2017 and https://github.com/PyCon/2017-slides"

AMLD2018 - Kevin Schawinski, ETH Zurich: Exploring the universe with AI
The Applied Machine Learning Days channel features talks and performances from the Applied Machine Learning Days. AMLD is one of the largest machine learning & AI events in Europe, focused specifically on the applications of machine learning and AI, making it particularly interesting to industry and academia.

How SpaceX Uses AI to Land their Rockets, Part I
If you wish to support us in creating fun, informative content, please consider giving at our Patreon site here: https://www.patreon.com/DrKnowItAllKnows. Thank you! Wherein Dr. Know-it-all begins examining how SpaceX actually lands their boosters, and their test vehicles like the Starship serial number 5 vehicle. What are the steps, and where can artificial intelligence (AI) help in the process? Watch to find out!

Artificial Intelligence In Space Exploration
In this video we will learn about the use of artificial intelligence in space exploration. Space is fascinating. There is hardly anyone who hasn’t looked at the night sky and imagined traveling through the space. There are hardly few rivalries that benefited mankind. One of them is the cold war between Russia and USA. That opened up the opportunity in space exploration. When one of these countries tried to shine over the other one, mankind shined with knowledge about space. And from there to today’s SPACEX age we have come across a long way. Starting from manually calculating the trajectory for moon landing to modern day usage of artificial intelligence in space exploration we have come a long way and yet there are many things to be done. By the way, did you know, Katherine Johnston, an African-American female physicist was the first person to do the calculations for the first actual moon landing in 1969.

Dark Matter

Youtube search... ...Google search ...News search

Machine Learning and Cosmology: New Approaches to Constraining the Dark Universe
Matias Carrasco Kind is a current astronomy graduate student and a CSE fellow at the University of Illinois at Urbana-Champaign. His current research interests lie in cosmology and extragalactic astronomy, especially in large scale structure, galaxy formation and evolution, computational and theoretical cosmology, environmental dependence of galaxy properties, photometric redshift estimation, machine learning techniques and data mining.

Dr Francois Lanusse - Exploring the Cosmos with Deep Learning
The main challenge of modern Cosmology is to answer pressing questions on the physical nature of dark matter and dark energy, which despite accounting for ~95% of the Universe today remain complete mysteries. This is what motivates a new generation of cosmological surveys which will map the Universe in great detail and on an unprecedented scale, implying a great potential for new discoveries but also new and outstanding challenges at every step of the science analysis, from image processing to the modelling of galaxy physics. In this talk, I will illustrate how recent advances in Deep Learning can be used to address some of these challenges and to exploit this wealth of data in new and exciting ways. A first application is the automated detection of rare astronomical objects, in this case strong gravitational lenses, using deep residual networks. In this typical image classification problem, Deep Learning has the potential of eliminating the need for human visual inspection which would have been intractable at the scale of future surveys. In a second example of application, I will present recent deep generative models and how they can be applied to emulate realistic signals when a proper physical model is lacking. In our specific application, we use deep generative models to produce realistic galaxy images, an essential part of the simulation pipeline necessary to the validation and calibration of our measurements. The last example I will mention is Deep Learning on graphs, more specifically how we use graph convolutional networks to model the properties of galaxies along the cosmic web (the large-scale structure of the Universe) in large-scale cosmological simulations.

Prof. Ofer Lahav: From Deep Learning to the Dark Universe
The talk will illustrate how new Artificial Intelligence and Machine Learning methods can help unveiling the mysteries of Dark Matter and Dark Energy from large deep galaxy surveys. Prof Ofer Lahav is an astrophysicist of international renown and holds the Perren Chair of Astronomy at University College London. He is famous for his advanced research on Dark Matter and Dark Energy by using methods like Machine Learning and is one of the founders of the Dark Energy Survey (DES). He has served as Vice-President of the Royal Astronomical Society and Co-Chair of the international DES Science Committee and has been awarded a wide variety of prices and honours. His projects include HST CLASH, DESI, LSST, Euclid, and several others focussed on understanding and analysing the behaviour of Dark Matter.

Galaxy Evolution

Youtube search... ...Google search ...News search

Understanding Galaxy Evolution through Machine Learning
Presentation for Summer All Zoom Epoch of Reionization Astronomy Conference (SAZERAC) 2020

John Wu - Deep learning in astrophysics: galaxy scaling relations
Deep learning in astrophysics: galaxy scaling relations. I will discuss some applications deep convolutional neural networks (convnets), and present a high-level overview of how to select, optimize, and interpret convnet models. We have trained a convnet to recognize a galaxy's chemical abundance using only an image of the galaxy; the traditional approach using spectroscopy requires at least an order of magnitude more telescope time and achieves a comparable level of accuracy to our method. We discover that the convnet can recover an empirically known scaling relation that connects galaxies' chemical enrichment and star formation histories with zero additional scatter, implying that there exists (and that the convnet has learned) a novel representation of the chemical abundance that is strongly linked to its optical-wavelength morphology.

Galaxies / Stars

Youtube search... ...Google search ...News search

Yashar Hezaveh: Mapping distant galaxies with artificial intelligence
We are in the midst of a revolution in computing, with "machine learning" algorithms solving problems for which major progress was thought to be decades away. As a result, computers can now exceed human performance at tasks such as recognizing patterns in images, playing complex games such as Go, and even driving cars. Can we harness the same machine learning algorithms to advance our understanding of physics? Understanding materials at the atomic level could lead to rapid advances in areas such as solar energy harvesting and the design of new computer hardware. The complex behavior of trillions of atoms all obeying the esoteric rules of quantum mechanics is not, however, easy for humans to untangle on their own. At this teachers' conference, we hear from scientists using the techniques of artificial intelligence and machine learning to advance our understanding of the quantum behavior of such matter as magnets, metals and superconductors, as well as the constituents of atoms. We also hear how insights from physics could be repurposed to advance the field of machine learning itself.

AI in the Sky | Artificial Intelligence in Space Exploration | upGrad - Future Forward
Right from the beginning of time, we have been studying and exploring outer space. Today, as we are receiving the most advanced data, Artificial Intelligence is contributing significantly to space exploration by helping scientists make empirical, smart decisions faster. upGrad presents #FutureForward, AI in Sky. Check out this video to know how AI is empowering scientists in the fields of astronomy.

Object Classification

Youtube search... ...Google search ...News search

Galaxy Classification with Machine Learning - Jim Geach
(University of Hertfordshire) on 08/03/2019.

Machine Learning Based Morphological Classification of Type Ia Supernova Host Galaxies
Gavin McCabe (Caltech) CASSI Symposium, 08/21/2020

Data, Science: Investigating 1 Million Galaxies with Humans and TensorFlow - Mike Walmsley
Deep learning is making astrophysicists excited and afraid. We're excited because we can show a neural network every galaxy in the sky and ask questions, and afraid because we don't understand the answers. I'll talk about how you can investigate galaxy evolution with TensorFlow and crowdsourced labels, what we're discovering, and where it all falls apart. www.pydata.org

Machine Learning Application on Galaxy Morphological Classification Using Dark Energy Survey Images
by Ting-Yun Cheng (University of Nottingham) on 08/03/2019.

CONVOLUTED COSMOS : Automatic Classification of Galaxy Images Using Deep Learning
In this project, Galaxy Image Classification using a Deep Convolution Neural Network is presented. The galaxy can be classified based on its features into three main categories, namely: Elliptical, Spiral, and Irregular. The proposed deep galaxy architecture consists of one input convolutional layer having 16 filters, followed by 3 hidden layers, 1 penultimate dense layer and an Output Softmax layer. It is trained over 3232 images for 200 epochs and achieved a testing accuracy 97.38% which outperformed conventional classifiers like Support Vector Machine and previous research contributions in the same domain of Galaxy Image Classification.

Astronomical Source Classification with Deep Learning
This video is a 2 minute introduction to a project (with a longer video here: https://youtu.be/kRuVtPhhKJU) that uses deep learning and convolutional neural networks to classify X-Ray images from the Chandra X-Ray Observatory into classes of either galaxy or galaxy cluster. It was developed as part of Harvard E89 Deep Learning Final Project.

Black Holes

Youtube search... ...Google search ...News search

Sensing to Intelligence: From Human Body to Black Holes in Space / Caltech Seminar Day Session II
Caltech professors Azita Emami and Katherine Bouman discussed a holistic approach that brings together the two distinct fields of sensor technology and artificial intelligence through the co-design of sensors and algorithms. The aim is to create sensor networks that intelligently and autonomously process their own data and adapt their activity in real time based on what they detect, enabling transformative discoveries in areas ranging from neural signals in the human body to black holes in space.

Speakers: Katherine Bouman Assistant Professor of Computing and Mathematical Sciences and Electrical Engineering; Rosenberg Scholar

Azita Emami Andrew and Peggy Cherng Professor of Electrical Engineering and Medical Engineering; Investigator, Heritage Medical Research Institute; Executive Officer for Electrical Engineering

What Would a Journey to the Black Hole Be Like?
Today, I'm setting off on a journey toward the nearest black hole. But don't worry - I'll keep you in the know by live-streaming my entire adventure! I'll have someone to talk to during the flight, and he can help me if things get really tough! My travel buddy's name is Liam. Liam is a robot with artificial intelligence. Space distances are seriously long. That's why traveling there would take way more time than you'd like to spend on the road! For example, Voyager 1, a space probe launched in 1977, was traveling out of the Solar System at a speed of 40,000 miles per hour. If my spacecraft moved at the same speed, it would take me a whole 77,000 years to get to the nearest star! But luckily, my spaceship is much faster than that. So let the journey begin!

Sun / Solar

Youtube search... ...Google search ...News search

Unity Alpha 0.03 - Enhanced AI Tracking, Solar System, Sun
Implemented a basic solar system, AI tracking that is more accurate and aggressive in how it will target you. Also tested some concepts fro sun designs (this video shows multiple suns). There are some bugs to work out with camera clipping near the larger sun, and also the AI's movement is still quite rigid, I'd like to make it more fluid in the future.

The Path to Safe Fusion Energy through Deep Learning
When the plasma inside a fusion reactor becomes unstable, there can be a release of energy that seriously damages the reactor. Key to preventing damage is predicting when these disruptions are about to happen. Professor William Tang of Princeton has successfully created a neural network that predicts these disruptions to a high level of accuracy, and was invited to give a talk at SDSC on his work. Tang's neural network is tuned specifically to predict disruptions, but the techniques used to train the neural network have crosscutting applications to other areas of scientific interest like cancer research.

Planets

Youtube search... ...Google search ...News search

Sara Seager: Search for Planets and Life Outside Our Solar System | Lex Fridman Podcast #116
Sara Seager is a planetary scientist at MIT, known for her work on the search for exoplanets.

Looking for exoplanets using supervised machine learning
"Data science in astronomical high-contrast imaging: looking for exoplanets using supervised machine learning" by Carlos Alberto Gomez Gonzalez , IPAG, Université Grenoble Alpes

Artificial Intelligence and NASA Data Used to Discover Eighth Planet Circling Distant Star
Our solar system now is tied for most number of planets around a single star, with the recent discovery of an eighth planet circling Kepler-90, a Sun-like star 2,545 light years from Earth. The planet was discovered in data from NASA’s Kepler space telescope. The newly-discovered Kepler-90i -- a sizzling hot, rocky planet that orbits its star once every 14.4 days -- was found by researchers from Google and The University of Texas at Austin using machine learning. Machine learning is an approach to artificial intelligence in which computers “learn.” In this case, computers learned to identify planets by finding in Kepler data instances where the telescope recorded signals from planets beyond our solar system, known as exoplanets.

AI helped NASA find a new planet!
Follow Nexter for more: https://nexter.org/ In the most recent NASA news, they have discovered an eighth planet which revolves around the star. This new planet adds to the growing Earth like planets list. Known as one of the Kepler planet, this planet was discovered by Google’s artificial intelligence by thoroughly searching those weak signals which has been overlooked by the Kepler Space Telescope. This newest galactic figure adds up the new planets discovered by NASA. This new planet, which is being touted as “Earth 2.0”, this Kepler new planet is located around Kepler-90 which is a star which has a distance of 2,500 light years from earth. This star was discovered in 2014. The new planet, however, cannot handle life like the Earth does because of its solar system. Kepler 90 is bigger by 20 percent and 5 percent hotter compared to our sun. The participation of computers and artificial intelligence in the planet searching process should not be of concern, however, by human beings. Artificial intelligence and technology are just there to help scientists and researchers look for more planets and find one that can support human life. More and more Earth sized planets are being discovered and being added into the NASA new planets. A lot of them are Kepler exoplanets. Astronomy and heavenly bodies are something that is very interesting, and every intergalactic news is being followed instantly. For more news and fun facts about the outer space, check out Nexter now.

SETI Talks - Big Astronomy Begins: Searching for Exoplanets with AI
To uncover the mysteries of the universe, astronomers are becoming greedy, making more observations than they can possibly analyze manually. Large photometric surveys from space telescopes like Kepler and the future TESS are no exception and today modern astronomers use artificial Intelligence (AI) algorithms to help them reveal the existence of exoplanets hidden in many years of observations of hundreds of thousands of stars. For this SETI Talk, we invited two researchers involved in the Kepler mission and AI to discuss the potential of neural networks to transform astronomy. Jeff Smith, Data scientist at the SETI Institute, has developed data processing and planet detection algorithms for Kepler since 2010 and is now involved in developing the pipeline for the future TESS mission. Chris Shallue, a senior software engineer at Google AI has used a neural network to analyze archival data from the Kepler Space Telescope to reveal the existence of two unknown exoplanets, named Kepler-90i and Kepler-80g. After presenting their recent work, we will discuss the impact of this new mode of scientific discovery, where artificial intelligence can assist humans in mapping out parts of the galaxy that have not yet been fully revealed.

Chris Shallue is a Senior Research Software Engineer on the Google AI team in Mountain View, California. His research is currently focused on machine learning techniques for identifying planets in data collected by the NASA Kepler space telescope. He also works on image captioning, natural language modeling and machine learning theory. Chris was previously a member of the Google Display Ads team where he worked on ad selection and personalization for GMail and Google Maps. Prior to joining Google, Chris was teaching, studying and researching mathematics.

Dr. Jeffrey Smith began his academic passion in the field of High Energy Accelerator Physics. His Ph.D. thesis was on the design of the International Linear Collider (ILC), a 22 mile-long electron-positron accelerator that will complement the discoveries being made at the Large Hadron Collider (LHC) at CERN in Geneva, Switzerland. After Cornell, Jeff joined the SLAC National Accelerator Laboratory at Stanford University to continue his work on the ILC and also to develop upgrade hardware for the LHC. After a successful career looking into the tiniest of inner-spaces Jeff decided to look up to the stars. Dr. Smith, now at the SETI Institute, develops data processing and planet detection algorithms for the Kepler and TESS Missions. Eking out planet signals in the Kepler data has proven to be a challenging and rewarding endeavor but looking toward the future, Dr. Smith is involved with developing new methods for use with the Transiting Exoplanet Survey Satellite (TESS), a new NASA planet finding mission to find Earth’s nearest cousins in a our galactic back yard.

Astrobiology and the Search for Extraterrestrial Life - with Ian Crawford
What can modern results in astrobiology tell us about the prospects for finding intelligent life elsewhere in the Universe? The famous Drake equation, which provides a rough estimate of the number of civilisations in our galaxy, predicts that space should be teaming with aliens. So where are they and why have we not found them yet? Watch the Q&A here: https://youtu.be/l7Xh_aphD30 Ian Andrew Crawford is professor of planetary science and astrobiology at Birkbeck, University of London. Crawford is a specialist in the science and exploration of the Moon and in the search for life in the Universe. Before switching his research interests to planetary science in 2003, Crawford had a 15-year career at University College London as an observational astronomer specialising in studies of the interstellar medium. He is the author of over 130 peer-reviewed research papers in the fields of astronomy, planetary science, astrobiology and space exploration. Crawford is a Fellow, and currently Vice President, of the Royal Astronomical Society, and a former member of the European Space Sciences Committee (ESSC) of the European Science Foundation. In 2014 he was appointed to the European Space Agency's Human Spaceflight and Exploration Science Advisory Committee (HESAC).

Earth

Youtube search... ...Google search ...News search

Microsoft's AI For Earth
Human society is faced with an unprecedented challenge to mitigate and adapt to changing climates, ensure resilient water supplies, sustainably feed a population of 10 billion, and stem a catastrophic loss of biodiversity. Time is too short, and resources too thin, to achieve these outcomes without the exponential power and assistance of AI. Early efforts are encouraging, but current solutions are typically one-off attempts that require significant engineering beyond what’s available from the AI research community. In this session we will explore, in collaboration with the Computational Sustainability Network (a twice-funded National Science Foundation (NSF) Expedition) the latest applications of AI research to sustainability challenges, as well as ways to streamline environmental applications of AI so they can work with traditional academic programs. The speakers in this session will set the scene on the state of the art in AI for Earth research and frame the agenda for the next generation of AI applications.

Neal Jean, " "Combining satellite imagery and machine learning to predict poverty"
Neal Jean, Michael Xie, Stefano Ermon, Matt Davis, Marshall Burke, David Lobell "Combining satellite imagery and machine learning to predict poverty" Stanford University Depts of Computer Science and Earth Systems Science CompSust-2016 4th International Conference on Computational Sustainability July 7, 2016

Geospatial Machine Learning for Urban Development
Ilke Demir, Postdoctoral Research Scientist, Facebook Presented at MLconf 2018 Abstract: The collective mission of mapping the world is never complete: We need to discover and classify roads, settlements, land types, landmarks, and addresses. The recent proliferation of remote sensing data (overhang images, LiDAR, sensors) enabled automatic extraction of such structures to better understand our world. In this talk, we will first mention the motivation and results of DeepGlobe Satellite Image Challenge[1][2], for road extraction, building detection, and land cover classification. Then we will go into details of an example approach[3] which proposes a complete system to use deep learning for generating street addresses for unmapped developing countries. The approach applies deep learning to extract road vectors from satellite images, then processes the street network to output linear and hierarchical street addresses, by labeling regions, roads, and blocks; based on addressing schemes around the world and coherent with human cognitive system. We will share and demonstrate the motivation and algorithm behind the scenes, then compare them to current open and industrial solutions, and walk through our open source code[4] to generate the addresses for a given bounding box.

Machine Learning with Earth Observation Imagery
AWS Public Sector Summit 2018 - Washington, D.C. For just a moment, think of the immense amount of data generated by Earth-observing systems. The sheer volume often makes it impractical for humans alone to perform the analysis, and accordingly, many groups are turning to artificial intelligence (AI) and machine learning (ML) algorithms to support their analysis. We'll hear from Development Seed and EOS about how they are using AI and ML to unlock the power of this planetary-scale data that is becoming increasingly more accessible in the cloud. From open-source libraries and human-in-the loop initial processing passes, to fully automated pipelines, we'll examine the new capacity for analysis now possible with technology.

Mars

Youtube search... ...Google search ...News search

The 'Space Architects' of Mars | The Age of A.I.
Life on Mars won't be how David Bowie pictured it. With the help of artificial intelligence, architects and scientists are painting a picture of what life in space could be like and how it will be possible. The Age of A.I. is a 8 part documentary series hosted by Robert Downey Jr. covering the ways Artificial Intelligence, Machine Learning and Neural Networks will change the world.

STEAM | Using AI to build our best possible future on Mars
Astreia Founder Dr. Natalie Rens joins us to talk about Artificial Intelligence. We cover her plans to use AI here on Earth and how that AI can then assist on colonizing the Red Planet and beyond.

Moon

Youtube search... ...Google search ...News search

AI and exploration - the Moon, astronaut health, Starspots and beyond
In this episode of SETI Live we’ll be looking at how AI is coming of age as a tool for space exploration. Space engineering is notoriously exacting, however AI’s ability to make predictions, mine data for insight, streamline workflows and inform decisions is opening up a revolution in discovery ‘down-stream.’ Indeed, as NASA's data gathering infrastructure scales into the peta-byte era, AI will become a crucial tool in our ability to make sense of the vast amounts of data being collected. This session will discuss the implications and opportunities being unlocked by AI for space exploration, with FDL partners from Google Cloud, Intel and the Luxembourg Space Agency - but also the potential for terrestrial applications informed by space research, such as understanding the causal processes of cancer, with the Mayo Clinic.

Data, Science: Investigating 1 Million Galaxies with Humans and TensorFlow - Mike Walmsley
Deep learning is making astrophysicists excited and afraid. We're excited because we can show a neural network every galaxy in the sky and ask questions, and afraid because we don't understand the answers. I'll talk about how you can investigate galaxy evolution with TensorFlow and crowdsourced labels, what we're discovering, and where it all falls apart. www.pydata.org

Artificial Intelligence: Powering Human Exploration of the Moon and Mars (Jeremy Frank; NASA AMES)
Talk by Dr. Jeremy Frank of NASA AMES at Arizona State University. (A joint FSE-SESE Colloquium co-hosted by CIDSE, SESE and SEMTE) 2/18/2020 Over the past decade, the NASA Autonomous Systems and Operations (ASO) project has developed and demonstrated numerous autonomy enabling technologies employing AI techniques. Our work has employed AI in three distinct ways to enable autonomous mission operations capabilities. Crew Autonomy gives astronauts tools to assist in the performance of each of these mission operations functions. Vehicle System Management uses AI techniques to turn the astronaut's spacecraft into a robot, allowing it to operate when astronauts are not present, or to reduce astronaut workload. AI technology also enables Autonomous Robots as crew assistants or proxies when the crew are not present. We first describe human spaceflight mission operations capabilities. We then describe the ASO project, and the development and demonstration performed by ASO since 2011. We will describe the AI techniques behind each of these demonstrations, which include a variety of symbolic automated reasoning and machine learning based approaches. Finally, we conclude with an assessment of future development needs for AI to enable NASA's future Exploration missions.

lec 47 artificial moon,artificial intelligence,neural network,deep learning
futurecop IAS


Man-made (artificial) Satellites

Youtube search... ...Google search ...News search

(A25) Artificial Intelligence in Space: Change Detection with Radar Satellite Data
ICEYE revolutionized the application of conventional radar satellite imaging, significantly bringing down the cost of launching small radar satellite missions. Learn how ICEYE has enabled broader use cases including using the technology to track pirates on the open seas. Session Speakers: Pekka Laurila

Katherine Scott: Machine Learning and AI for Satellite Imaging
At the 2017 Hackaday Superconference Katherine Scott sat down with Hackaday to talk about her work at Planet Labs. As Image Analytics Team Lead she stitched together 1.2 Million images taken by cubesats to get a daily picture of the earth, then set about using machine learning, computer vision, and AI to make sense of the changing data.

When deep learning meets satellite imagery
A handy guide to understanding the specificities and challenges of satellite images when using deep learning. Speaker - Julie Imbert Script - Julie Imbert and Ségolène Husson With the participation of Renaud, Tugdual, Thomas and all the Earthcube team

Landuse Classification from Satellite Imagery using Deep Learning
With the abundance of remote sensing satellite imagery, the possibilities are endless as to the kind of insights that can be derived from them. One such use is to determine land use for agriculture and non-agricultural purposes. In this talk, we’ll be looking at leveraging Sentinel-2 satellite imagery data along with OpenStreetMap labels to be able to classify land use as agricultural or non-agricultural. Sentinel-2 data has a 10-meter resolution in RGB bands and is well-suited for land use classification. Using these two datasets, many different machine learning tasks can be performed like image segmentation into two classes (farm land and non-farm land) or more challenging task of identification of crop type being cultivated on fields. For this talk, we’ll be looking at leveraging convolutional neural networks (CNNs) built with Apache MXNet to train deep learning models for land use classification. We’ll be covering the different deep learning architectures considered for this particular use case along with the appropriate metrics. We’ll be leveraging streaming pipelines built on Apache Flink and Apache NiFi for model training and inference. Developers will come away with a better understanding of how to analyze satellite imagery and the different deep learning architectures along with their pros/cons when analyzing satellite imagery for land use.

Asteroids

Youtube search... ...Google search ...News search

How machine learning helps scientists track asteroids
When NASA issued a worldwide challenge to help them better track the asteroids and comets that surround Earth, Gema Parreño answered the call. She used #TensorFlow, Google’s machine learning tool, to create a program called Deep Asteroid, which helps identify and track Near Earth Objects. Special thanks to the Royal Observatory of Madrid.

How to avoid the Asteroid? | Deep Asteroid
Deep asteriod : Predictive model of NEOs’ trajectory using ‘Deep Learning’ and ‘TensorFlow’ Idea for NASA Space Challenge 2016

Collision Avoidance in Space

Youtube search... ...Google search ...News search


Challenge: Today, active collision avoidance among orbiting satellites has become a routine task in space operations, relying on validated, accurate and timely space surveillance data. For a typical satellite in Low Earth Orbit, hundreds of alerts are issued every week corresponding to possible close encounters between a satellite and another space object (in the form of conjunction data messages CDMs). After automatic processing and filtering, there remain about 2 actionable alerts per spacecraft and week, requiring detailed follow-up by an analyst. On average, at the European Space Agency, more than one collision avoidance manoeuvre is performed per satellite and year. In this challenge, you are tasked to build a model to predict the final collision risk estimate between a given satellite and a space object (e.g. another satellite, space debris, etc). To do so, you will have access to a database of real-world conjunction data messages (CDMs) carefully prepared at ESA. Learn more about the challenge and the data.


Results: Spacecraft collision avoidance procedures have become an essential part of satellite operations. Complex and constantly updated estimates of the collision risk between orbiting objects inform the various operators who can then plan risk mitigation measures. Such measures could be aided by the development of suitable machine learning models predicting, for example, the evolution of the collision risk in time. ...This paper describes the design and results of the competition and discusses the challenges and lessons learned when applying machine learning methods to this problem domain. Spacecraft Collision Avoidance Challenge: design and results of a machine learning competition | T. Uriot, D. Izzo, L. Simoes, R. Abay, N. Einecke, S. Rebhan, J. Martinez-Heras, F. Letizia, J. Siminski, and K. Merz

The Truth About Space Debris
Realengineering

This 19-year-old can keep astronauts safe from space junk
Space debris is a growing threat to space exploration. When 15-year-old Amber Yang first heard about space trash it gave her nightmares. How could such an imminent threat to space exploration be left on the back burner? After seeing videos with astronaut Scott Kelly, she decided to take matters into her own hands. Within a few years she consumed all the media she could on space debris, taught herself to code, and learned the ins and outs of astrophysics. By 18 she had developed an AI-based space debris tracking program that she claims is one of the most accurate in the world.


Astropy Project

Youtube search... ...Google search ...News search

  • Astropy Project ...a community effort to develop a common core package for Astronomy in Python and foster an ecosystem of interoperable astronomy packages.

Practical Applications of Astropy | SciPy 2018 | David Shupe
In this talk intended for the astronomy mini-symposium, I will highlight how capabilities in Astropy are used to solve computational problems in two current projects. Coordinate matching, world coordinate systems, and data manipulation are used heavily in parts of the data system for the Zwicky Transient Facility, which processes 1 terabyte of images per night, within 10 minutes of receipt of each exposure. Astropy units and quantities, and the cosmology module, enabled the rapid development of a web application for the proposed Origins Space Telescope.

PyAstro20: State of the Package: Astropy
Brett Morris gives an overview of the Astropy ecosystem

Astropy Beyond Astronomy: Infrastructure of an Open Source Ecosystem | SciPy 2019 | B. Sipocz
The Astropy Project is a community effort to develop a common core package and to foster an ecosystem of interoperable astronomy packages. In addition to these domain specific libraries we also maintain a suite of infrastructure packages. This infrastructure can be easily utilized to cover the needs of most scientific Python projects. Sharing the tools we developed for maintenance, testing, documenting, and general upkeep of the most essential parts of our project maximizes the impact of the software engineering expertise and resources within the whole ecosystem to ensure that others can benefit from the efforts of the core Astropy team.

Astropy and astronomical tools Part I | SciPy 2014 | Greenfield, Bray, Robitaille, Aldcroft
Enthought

Simulation

Youtube search... ...Google search ...News search

Powerful A.I. Engine simulates the Universe Formation
For the first time, scientists have used artificial intelligence to create complex, three-dimensional simulations of the Universe. It's called the Deep Density Displacement Model, or (D^3)M, and it's so fast and so accurate that the astrophysicists who designed it don't even know how it does what it does. What it does is accurately simulate the way gravity shapes the Universe over billions of years. Each simulation takes just 30 milliseconds - compared to the minutes it takes other simulations.

Computing a Universe Simulation
Physics seems to be telling us that it’s possible to simulate the entire universe on a computer smaller than the universe.

Shirley Ho - The First AI Simulation of the Universe is Fast and Accurate (February 26, 2020)
A full understanding of the evolution of the universe’s structure is one of the holy grails of modern astrophysics. Astrophysicists survey large volumes of the universe and compare the findings to computer simulations. Simulating the movement of billions of particles over billions of years is a daunting task, however, even when using the simplest physical models. In this lecture, Shirley Ho will discuss her team’s work building a deep neural network that learns from a set of pre-run numerical simulations and predicts the large scale structure of the universe. Extensive analysis demonstrates that their deep-learning technique outperforms the commonly used fast approximate simulation method in predicting cosmic structure in the non-linear regime. They also show that their method can accurately extrapolate far beyond its training data and predict structure formation for significantly different cosmological parameters. This ability to extrapolate outside its training set is highly unexpected and remains a mystery.

Deep Generative Modeling for HEP, Cosmology and Fluid Dynamics
Karthik Kashinath of NERSC discusses using deep generative modeling for HEP, cosmology and fluid dynamics

AI Learning to land on The Moon. Lunar Lander V2 reinforcement learning
This deep learning reinforcement algorithm is using an off policy Sarsa (State-Action-Reward-State-Action) Agent to learn to control this moon landing capsule.The AI can use the left thrust, right thrust, main thrust or do nothing. This is running agent is running in OpenAI Lab. OpenAI Lab is an experimentation framework for Reinforcement Learning using OpenAI Gym, TensorFlow, and Keras. With Lunar Lander V2 the agent gets a reward for moving from the highest point of the screen to landing zone and landing with zero speed. On the off chance that lander moves far from landing cushion it loses reward points. Episode completes if the lander crashes or lands, getting extra points. Using main thruster cost points. Fuel is endless, so the agent can figure out how to fly and then land. The data on the right shows reward overtime with our 150 episode exploration rate, mean reward overtime and the network loss overtime.

Landing a SpaceX Falcon Heavy Rocket
Can we land a SpaceX Falcon Heavy Rocket in simulation using machine learning? Yes! Reinforcement learning is a technique that lets an agent learn how best to act in an environment using rewards as its signal. OpenAI released a library called Gym that lets us train AI agents really easily. We'll use a combination of the TensorFlow and gym libraries to build an RL agent capable of landing a rocket perfectly. The specific technique we're using is called proximal policy optimization, this is an actor-critic algorithm that is really popular. Lets get started!