Computational Training
TAP runs Hackathons and Computational Workshops to ensure exposure to state-of-the-art computational methods (machine learning/AI) and training in how to use astrophysical codes that drive cutting-edge research in Theoretical Astrophysics.
Upcoming Events
Fall 2025 Computation & Data Initiative Workshops
Dalya Baron, Stanford University
Computation & Data Initiative Workshops
Date: TBA
Title: TBA
Fall 2025 Planet Formation Initiative Computational Workshops
Cristiano Longarini, Cambridge University
Planet Formation Initiative Computational Workshop
Title: TBA
Date: TBA
Previous Events
Computation & Data Initiative Computational Training
Hack Arizona Event
March 22, 2025
TAP Computation & Data Initiative hosted three Hack Arizona sessions
The IAU 2024 General Assembly in Capetown, South Africa
August 2024
TAP, a partner for the IAU 2024 General Assembly held in Capetown, South Africa, was featured in the event’s poster sessions and event website. This was an exceptional opportunity to be in front of a large global audience of potential faculty, postdocs, and graduate students. TAP Faculty Member Chi-Kwan Chan represented TAP with a project poster submission and Q&A.
Deep Learning for Astrophysics Hackathons
Spring 2024
Organized under the TAP Computation and Data Initiative, Chi-Kwan Chan ran weekly hackathons (11 sessions) open to all TAP members in FY24. 18 participants from Astronomy, Physics, and Computer Science joined the hackathons. The sessions were designed to teach a variety of applications for deep learning in science, train on neural network learning, and provide hands-on experience working through challenging projects and solutions.
Overview doc of all the sessions
Links used for Hackathon Teaching Concepts:
Cosmology Initiative Computational Training
Practical Session on Using Generative Models
May 5, 2025
Carol Cuesta-Lazaro, Massachusetts Institute of Technology
TAP Cosmology Initiative Computational Workshop II
Speaker encourages everyone to use their own research.
BIO: Carol is an IAIFI Fellow at the NSF Institute for Artificial Intelligence and Fundamental Interactions. She splits her time between MIT and the Center for Astrophysics at Harvard, developing robust machine learning models that can guide us towards future discoveries in physics. Carolina received her PhD in Physics and Data Intensive Science from the Institute of Computational Cosmology at Durham University, UK, in 2022. Alongside her PhD, she has been a research collaborator with the United Nations (UN) Global Pulse and the UK’s National Health Service (NHS), developing epidemiological simulations, and a research intern at Amazon’s Alexa team. She is also passionate about science communication and making astrobabble understandable.
From Zero to Generative: Learning Generative Models from Scratch
April 29, 2025
Carol Cuesta-Lazaro, Massachusetts Institute of Technology
TAP Cosmology Initiative Computational Workshop I
This workshop introduces generative modeling with no prerequisites in machine learning or deep learning. Starting with basic concepts, we guide participants through the intuition and implementation of key generative frameworks, such as diffusion models. Participants will gain hands-on experience building simple versions of these models and understanding when to apply different architectures.
BIO: Carol is an IAIFI Fellow at the NSF Institute for Artificial Intelligence and Fundamental Interactions. She splits her time between MIT and the Center for Astrophysics at Harvard, developing robust machine learning models that can guide us towards future discoveries in physics. Carolina received her PhD in Physics and Data Intensive Science from the Institute of Computational Cosmology at Durham University, UK, in 2022. Alongside her PhD, she has been a research collaborator with the United Nations (UN) Global Pulse and the UK’s National Health Service (NHS), developing epidemiological simulations, and a research intern at Amazon’s Alexa team. She is also passionate about science communication and making astrobabble understandable.
Dynamics Initiative Computational Training
Orbital Dynamics in Python
March 18, 2025
Daniel Tamayo, Harvey Mudd College
TAP Dynamics Initiative Computational Workshop
This computational workshop provided a hands-on introduction to the REBOUND N-body integration package. It introduced some of the basics, discussed its new visualization tools, as well as going through how to customize the code for your own scientific goals. We showed how various common astrophysical effects (tides, oblateness, radiation pressure etc.) can be easily incorporated using the REBOUNDx library, and how you can add your own custom effects. Finally, we introduced celmech, a new Python package for facilitating theoretical orbital mechanics calculations, and isolating the effects of resonances, secular dynamics, and much more!
Planet Formation Initiative Computational Training
Photochem Workshop II
November 5, 2024
Nicholas Wogan, NASA Ames Research Center
An Open-Source Chemical and Climate Model of Planetary Atmospheres
Learning Objectives:
A brief introduction to photochemistry and climate modeling
Installing Photochem
First photochemistry simulation: a model of Modern Earth
Modifying chemical reaction rates or other internal model data (e.g., photolysis or aerosol cross sections)
Modifying the Photochem source code & contributing to the model
First climate simulations: reproducing the Habitable Zone
If we have time – a photochemistry and climate simulation of TRAPPIST-1e assuming it is an inhabited world.
Photochem Workshop I
November 4, 2024
Nicholas Wogan, NASA Ames Research Center
An Open-Source Chemical and Climate Model of Planetary Atmospheres
Learning Objectives:
A brief introduction to photochemistry and climate modeling
Installing Photochem
First photochemistry simulation: a model of Modern Earth
Modifying chemical reaction rates or other internal model data (e.g., photolysis or aerosol cross sections)
Modifying the Photochem source code & contributing to the model
First climate simulations: reproducing the Habitable Zone
If we have time – a photochemistry and climate simulation of TRAPPIST-1e assuming it is an inhabited world.
Computational Training Resources
High Performance Computing (HPC)
The University of Arizona offers free high performance computing resources in the Research Data Center (RDC), a state-of-the-art facility that hosts large computer clusters. Scientific research relies upon such systems to analyze massive amounts of collected data.
Benefits:
Several HPC clusters support different research computing needs whether it is for hundred’s of small jobs that only use a few cores each, or very large jobs using hundreds of cores or terabytes of memory. Or anything in between. There are also many nodes with Nvidia GPU’s.
Details are available at Compute Resources.
Service Offerings:
- A web interface to the cluster offers a file browser, a graphical desktop, and the Jupyter notebooks and RStudio applications
- An account management portal which lets you manage your account and request services
- Workshops and training in a range of HPC subjects from Introductory HPC to MPI
- The HPC is not compliant for export-controlled or CUI data