Big Data from Space 2021
18-20 May 2021 | Virtual Event
Keynote Speakers
Peter Bauer
ECMWF
Keynote Lecture: Tuesday, 18 May 2021 - 11:00
Title - Digital twins of the Earth system = really big data
Abstract - Digital twins aim to deliver a view of the changing Earth with unprecedented detail and accuracy allowing to monitor and predict change. They will serve the preparation for extreme events, the mitigation of their impacts and help devise adaptation strategies for all aspects of society. Digital twins require bringing together data from high-resolution simulation models and a vast fleet of satellite observations complemented by ground based networks and novel data sources like the internet of things.
The volume and diversity of data will drastically increase, and this has implications on data networks, computing infrastructures and data pre- and post-processing workflows. The challenges of the technical elements of data handling and value adding across federated infrastructures need to be met aiming to facilitating access to data and offering new data analytics capabilities at the same time. The European Commission Destination Earth project offers a unique opportunity to meet these challenges and deliver a lasting European infrastructure for the benefit of European society.
Fernando Pérez
UC Berkeley
Keynote Lecture: Tuesday, 18 May 2021 - 16:00
Title - From Interactive Computing To Collaborative Science:Opportunities In The Cloud With Open Infrastructure
Abstract - Open source has "won the race" in science, in the sense that today virtually all major scientific projects (as well as everyday individual research and teaching) are powered to some extent by open source tools. The next revolution in scientific practice is already apparent: cloud infrastructure, powered by open source tools, give us the promise of a new "data watering hole" where scientific communities can grow. Vast amounts of data that cover multiple aspects of a given problem, ready to be analyzed with open tools, offer the potential for rich interdisciplinary practices where teams combine multiple data sources and modeling/analysis approaches.
To realize this vision, we need our open tools to be developed sustainably, to be recognized as a critical part of the scientific enterprise, and to be governed in a multi-stakeholder fashion that empowers scientists and educators as first-class partners of the conversation and not only as consumers of commercial cloud offerings. But the commercial cloud vendors also have a critical role to play in this space, as their infrastructure is unmatched by anything that even large scientific collaborations can hope to achieve.
I will discuss some of this vision, centered around open source projects like Jupyter and Pangeo that aim to provide scientists with the tools, practices and communities to make it a reality. I will highlight some of the technical elements in these projects as well as discuss some of the broader challenges that exist regarding incentives, careers and stakeholders.
Markus Reichstein
MPG, Jena
Keynote Lecture: Wednesday, 19 May 2021 - 11:00
Title - Earth Observation + Machine Learning + System Modelling To Understand The Earth System
Abstract - The Earth is a complex dynamic networked system. Machine learning, i.e. derivation of computational models from data, has already made important contributions to predict and understand components of the Earth system, specifically in climate, remote sensing and environmental sciences. For instance, classifications of land cover types, prediction of land-atmosphere and ocean-atmosphere exchange, or detection of extreme events have greatly benefited from such approaches which centrally included Earth Observation. Such observation-driven information has already changed how Earth system models are evaluated and further developed. However, many studies have not yet sufficiently addressed and exploited dynamic aspects of systems, such as memory effects for prediction and effects of spatial context, e.g. for classification and change detection. In particular new developments in deep learning have shown great potential to overcome these limitations.
Radu Ionicioiu
IFIN-HH, Bucharest
Keynote Lecture: Wednesday, 19 May 2021 - 14:00
Title - Space: The Quantum Frontier
Abstract - Quantum technologies are bringing a paradigm change in several fields, from computation and simulation to communication, sensing and imaging.
Genuine quantum properties, like superposition, entanglement and wave-particle duality are allowing us to perform things not possible in a classical world: teleportation, interaction-free imaging and efficient quantum algorithms, to name only a few. This justifies the mantra of the second quantum revolution"Quantum is a resource". Thus, using quantum-stuff, we can do better-stuff.
After an overview of this brave new quantum world, I will discuss future quantum applications in space technologies.
Leanne Guy
AURA/Rubin Observatory
Keynote Lecture: Thursday, 20 May 2021 - 09:00
Title - Big Data Astronomy: Challenges and Opportunities
Abstract - Driven by advances in telescope, sensor and computing technology, modern digital astronomical surveys such as the Sloan Digital Sky Survey (SDSS) and ESA's Gaia mission have produced massive amounts of data, enabling a very wide variety of scientific investigations. The Vera C. Rubin Observatory, scheduled to begin operations in late 2023, will execute as its prime mission the 10-year Legacy Survey of Space and Time (LSST). LSST will collect about 20 terabytes of multi-colour imaging data every night, culminating in a 500 petabyte set of images and data products after 10 years. The science enabled by Rubin's LSST will be very broad, ranging from studies of small moving bodies in the solar system to the structure and evolution of the universe as a whole. The LSST dataset is expected to profoundly affect the scientific landscape over the next ten years and beyond. In this keynote I will discuss how Rubin Observatory is preparing to manage of the massive stream of data generated by the LSST as well as novel approaches for extracting science from such a large and complex dataset.