Letters of Intent received in 2022

LoI 2024-2170
Rise of the Machines: Computational astrophysics opportunities and challenges of the coming decade

Topics

Numerical simulations
Computational methods for dealing with Big Data, data mining and processing
Machine Learning, Likelihood methods
Computational astrophysics training and resource sharing

Rationale

SOC:
Dmitry Bisikalo (Inst of Astronomy RAS)
Christian Boily (Observatoire astronomique) (co-Chair)
Pratika Dayal (University of Groningen)
Selma E. de Mink (Max Planck Institute for Astrophysics)
Stefan Ferreira (North West University)
Michiko Fujii (University of Tokyo)
Jonathan Mackey (Dublin Institute for Advanced Studies)
Shazrene Mohamed (University of Miami, South African Astronomical Observatory and University of Cape Town) (co-Chair)
Simon Portegies Zwart (Leiden University)
Russ Taylor (University of Cape Town, University of the Western Cape, IDIA)
Brian van Solen (University of the Free State)
Gustavo Yepes (Universidad Autonoma de Madrid)

The past decade has seen major advances in observational astronomy - the detection of gravitational waves with LIGO-Virgo, the start of large surveys e.g., Gaia, the successful deployment and commissioning of the JWST, and the Black Hole event horizon observations. At the same time, the African continent has seen explosive growth in multi-wavelength astronomy with the refurbishment of the SAAO telescopes, and the building of the 10m-class Southern African Large Telescope, and the MeerKAT radio telescope (64 dishes) in South Africa, as well as optical telescopes in Ethiopia (Entoto Observatory), the Bouzareah Observatory in Algeria, the refurbished Kottamia Astronomical Observatory in Egypt, Oukaimeden Observatory in Morocco, H.E.S.S. for TeV gamma rays in Namibia, and with optical telescopes planned in Burkina Faso, and Kenya (KOTI), and the construction of the world’s largest radio telescope, the SKA, already underway across the continent.

Using these advances in observational astronomy as a catalyst, now is the opportune moment to push for further development and growth on the computational side. To drive our understanding of astrophysical phenomena, we need to come up with more computationally efficient algorithms and approaches, more physically motivated, detailed simulations, and in fact, for African astronomers as well as the global astronomy community, such advances in numerical simulations and data handling will be critical to ensure the success of many of the forthcoming projects listed above.

A first major challenge is to ensure that we are able to deal with the imminent data deluge; in particular from the SKA and the Rubin Observatory. In this FM we will highlight the huge strides being made with respect to storing, processing, and analyzing Big Data on the African continent (e.g., with MeerKAT and ilifu, the data-intensive research cloud), but we also will discuss future approaches and possible solutions to known upcoming challenges. To prepare for these challenges, we have to develop the necessary computational infrastructure, including data processing centers and research clouds. Machine learning and AI is expected to play a prominent role in the near future, and its usefulness for astronomical research is growing rapidly. Preparing for the growing data, and the data complexity we have to develop new techniques in data mining, disseminating information and making the science tractable.

In connection with the latter, this coming era brings new challenges and opportunities for numerical simulations as well. In many cases, the current computational power has allowed us to move beyond low-resolution, simple, one-dimensional models: now with multi-dimensional simulations we strive to bridge vast dynamic ranges, in both space and time, in order to self-consistently capture essential micro- and macro-scale physical processes. Large grids of models with detailed physics, together with ever more sophisticated methods and computational algorithms, e.g., likelihood methods and/or machine learning are increasingly being used to interpret the observational data itself, e.g., the huge number of template models developed for fitting the EHT observations of M87 and Sgr A*. We will highlight the current state-of-the-art in different fields, both for the new approaches, as well as the more traditional simulations, e.g., solving the (GRM)HD equations and then often using the results to derive synthetic observables for comparison with real data. We will also discuss the infrastructure and training needed to grow the modelling expertise in Africa and other developing nations; such expertise is critical if we are to extract maximal benefit from the current and future facilities. Indeed, simulations have always been used to make predictions, but the ability now to combine them with instrument responses and data reduction pipelines has resulted in tools that can be used to guide the observational strategies themselves - bringing the observation/simulation cycle ‘full circle’.

The challenges also bring opportunities for innovation and importantly for the growth of the computational astrophysics community. This FM will draw-up a roadmap outlining the best ways to support this objective, building on the work already happening in many developing countries, the existing and planned access to computational infrastructure, and training, sharing information and resources. To this end, SOC members and invited speakers will also hold day-long post-meeting training sessions and advanced tutorials on techniques and community codes; these include the Astrophysical Multipurpose Software Environment (AMUSE), MESA (stellar evolution), (GR)MHD codes, the Machine Learning workshops (e.g., JEDI workshops), and Big Data training and hackathons (e.g., DARA). In partnership with the outreach teams at the South African Astronomical Observatory (SCBP) and South African Radio Astronomy Observatory (both headquartered in Cape Town), post-meeting visits to local schools in Cape Town (the majority of Western Cape schools are equipped with computer labs) will also be used to engage the broader community through outreach and education activities to demonstrate the use of computers in astronomy, this will include examples and exercises, talks about supercomputers and numerical simulations for advanced high school students, as well as teacher training workshops on tools for programming and simulations (e.g., SCRATCH which is translated into local languages, e.g., isiXhosa, isiZulu). Through these engagements we hope to expose a new generation to the exciting challenges and opportunities in computational astrophysics.