Vacation Studentships in Astronomy at CAS
The Centre for Astrophysics & Supercomputing (CAS) accepts applications for Vacation Studentships from enthusiastic university students with excellent scholastic records who are in the last, or second last, year of their undergraduate or Honours/Masters degree.
With 26 research faculty, 27 postdoctoral researchers, and nearly 50 PhD students, CAS is a vibrant, friendly environment for studying most fields of astronomy. Swinburne astronomers have guaranteed access to the twin Keck 10-m Telescopes in Hawaii - the world's premier optical observatory - and CAS owns and operates one of Australia's most powerful supercomputers - Ozstar. We also develop advanced, immersive 3D data visualization facilities and create 3-D animations and movies promoting and explaining astronomy to the broader community.
Swinburne's Hawthorn campus is situated in a lively, urban setting just minutes by public transport from Melbourne's city centre.
Our Vacation Studentship programme aims to provide undergraduate students with some insight into how exciting research is and how it is conducted. Students will join a research project, or possibly help start a new one, in one of the many areas of astronomy in which CAS staff and post-docs are experts. The various projects on offer are listed below. Projects can involve all aspects of astronomical research, from proposing or carrying out new telescope observations, to analysing data, to conducting theoretical calculations or advanced simulations. Many previous students have eventually published peer-reviewed research articles on some of their Vacation Studentship research.
In 2024 this programme is expected to run in-person, with remote options possible, and will be available to students currently enrolled at universities in Australia and New Zealand. Projects are expected to run over 8 weeks, between November and February, with the timing to be negotiated between the student and their nominated supervisor. Vacation students are paid a tax-free stipend of $750 per week. An additional $250 is available for students based in New Zealand to offset the increased cost of relocation.
Applications are now being accepted and should be received before
23rd August 2024; they should include the following:
Applications and reference letters should be emailed to Dr. Simon Stevenson (spstevenson@swin.edu.au) with the above information attached (preferably as PDF documents).
The cover letter is important and should:
(i) set out why you are interested in undertaking a vacation studentship at Swinburne and
(ii) list at least two research projects you are interested in working on (with an optional ranking). See below for the current list of projects on offer.
Potential Vacation Studentship Research Projects
The following list outlines particular projects currently on offer. Other projects not listed here may also arise. If you have questions, contact Dr. Simon Stevenson at the above email.
(Last Updated 26th July 2024)-
Preparing for the use of Large Language models in Astronomy - AstroChat
Astronomy surveys are collecting data at an unprecedented rate and we're relying on the use of machine learning and AI to help us sort and analysis data. With the introduction of large language models, like chatGPT, astronomers are starting to consider how these AI systems might be integrated into our everyday lives to help us accelerate discoveries. In this project, you'll work with current optical transient survey data to build up a training set of images and question items to train a custom Visual LLM at Swinburne for the use in transient astronomy. This will be a unique experience for you to learn about AI and LLMs, astronomical data and surveys, and work hands on with images from some of the largest astronomical facilities in the world.
Supervisor: Dr. Sara Webb
-
Understanding Ultra-Diffuse Galaxy Formation with the Keck Telescopes
The closest star to Earth (other than the Sun) is 4 lightyears away. If the Milky Way were an Ultra-Diffuse Galaxy it is likely the closest star would instead be more like 20 lightyears away. This is because Ultra-Diffuse Galaxies have 100x less stars than the Milky Way but are the same size. The night sky would look completely different. Understanding how galaxies can become so large, yet so poor at forming stars is an outstanding problem of current galaxy formation theory. This project uses spectroscopic data already observed on the 10m Keck telescopes to better understand Ultra-Diffuse Galaxy formation. The student will help analyse and understand these data using common astronomical data packages. A basic understanding of coding in Python is required.
Supervisors: Dr. Jonah Gannon and Prof. Duncan Forbes
-
Searching for Off-peak Emission from Gamma-ray Pulsars
The recent release of the The Third Fermi Large Area Telescope Catalog of Gamma-ray Pulsars (3PC) has increased the number of detected gamma-ray pulsars to almost 300. This provides a good number of detected gamma-ray pulsars to study various topics related to gamma-ray emission mechanisms of pulsars. A recent study (Song et al. 2023) indicates that pulsars might be emitting weak, isotropic gamma-ray emission. As a follow up to this study, this project is aiming to lay the foundation to search for this emission in detected gamma-ray pulsars when their rotational phase is off-peak. In this project, we will aim to perform the following tasks. We would first perform timing analysis of gamma-ray pulsars in 3PC using Tempo2 or PINT with existing timing solutions, and update them if necessary. We will then create an accessible table of on- and off-peak phases of each gamma-ray pulsar. If time allows, we will perform likelihood analysis on the off-peak phase of each individual pulsar.
Further Reading:
- The Third Fermi Large Area Telescope Catalog of Gamma-ray Pulsars (3PC): Smith et al. 2023
- Fermi-LAT Data Analysis with Fermipy: Fermipy
- Introduction to gamma-ray pulsar timing: Smith et al. 2008
Supervisors: Dr. Yuzhe (Robert) Song
-
Detecting Gamma-ray Emission from Stripped Envelope and Superluminous Supernovae
A recent stripped-envelope supernova, SN 2022jli, was observed to emit gamma-rays (Chen et al. 2024). In this study, the authors demonstrate the rare properties of the bound binary system that survived the supernova explosion, and provide insights into the formation and early evolution of X-ray binaries (XRBs). This can be in turn used as an input for population synthesis study of XRBs. In this project, we propose a stacking survey of stripped envelope supernovae (SESNe) and superluminous supernovae (SLSNe) in gamma-rays with Fermi-LAT data, similar to Renault-Tinacci et al. 2017 . We will use the Transient Name Server to first identify targets appropriate for this study, and perform individual and stacking analysis on the targets. If any detection is made, we will work to explain the physical implication of the results and use them to constrain population modelling of XRBs with COMPAS.
Further Reading:
- Fermi-LAT discovery of the GeV emission of the superluminous supernovae SN 2017egm: Li et al. 2024
- Fermi-LAT Data Analysis with Fermipy: Fermipy
- Search for gamma rays from SNe with a variable-size sliding-time-window analysis of the Fermi-LAT data: Prokhorov et al. 2021
Supervisors: Dr. Yuzhe (Robert) Song and Dr. Simon Stevenson
-
Axion Dark Matter Detection – Data Acquisition and Analysis
The nature of dark matter is one of the biggest mysteries in modern science – it makes up five sixths of the matter in the Universe, and is of unknown composition. It surrounds and passes through the Earth at all times. Axions are a hypothetical particle, and one of the leading candidates for dark matter. Swinburne is building a new axion detector to try and measure small effects induced by dark matter when it passes through the laboratory. The kind of experiment we are building is called an axion haloscope. The hope is that if the dark matter is made of axions, the detector can shed some light on its properties. The detector is being physically constructed and will be hosted at Swinburne – but work needs to be done on the software and analysis side.
This project will focus on a data acquisition and analysis pipeline for the new axion dark matter detector. You will be working on code to interface with laboratory equipment, acquire new experimental data, and then tease through that data looking for hints of new physics.
Supervisor: Dr. Ben McAllister
-
Are the fundamental constants of nature the same in distant galaxies?
Distant galaxies, seen in silhouette against bright, background quasars, imprint a characteristic pattern of absorption lines onto the quasar light as it travels to Earth. This pattern is determined by the fundamental constants of nature. Using spectra taken with the largest optical telescopes in the world (e.g. Keck and Subaru in Hawaii, VLT in Chile), this pattern can be compared with laboratory spectra to determine whether the fundamental constants were indeed the same in the distant, early universe as we measure them on Earth today. Several different avenues are available for exploration in this project. For example, one option is to analyse new spectra taken from the Keck and/or VLT with the aim of measuring the variability of the fine-structure constant (effectively, the strength of electromagnetism). Another option is to improve the methods used to make these exacting measurements so that we can make the best use of a new instrument on the VLT. These and other options will be discussed with the candidate.
Further Reading:
Supervisor: Prof. Michael Murphy
-
Machine learning classification of radio transients discovered by ASKAP telescope
Short bursts, flares, scintillation, and other radio time-domain phenomena usually imply extreme astrophysical environments (e.g., strong magnetic fields). Therefore, these objects, known as radio transients, can be used as laboratories to study extreme physics that cannot be studied on Earth. We developed the VASTER fast imaging pipeline to identify radio transients from the Australian Square Kilometre Array Pathfinder (ASKAP) telescope. The pipeline identified various transients, including pulsars, flaring stars, and extreme scintillating galaxies. However, it can also select false candidates that require manual inspection to rule them out. The aim of this project is to develop a machine learning-based classification algorithm for radio transients identified by the VASTER pipeline, determining whether they are real astrophysical objects or artefacts. Basic Python knowledge is preferred.
Further Reading:
Supervisor: Dr. Yuanming Wang
-
Studying the Nature of Fast Radio Bursts through their Energies
Fast Radio Bursts (FRBs) are bright, short flashes originating from distant galaxies. They have only been discovered in 2006, and it is still unknown what causes them. One important clue studied in detail is the energy distribution. Yet, one aspect is widely ignored, the directional dependence of the emission, i.e. the beaming. In this project, you will explore theoretically the influence of beaming on the energy function as far as possible. Beyond that, you will make your own simulation to see how beaming affects more complicated energy functions. Through trying of reproducing the observed energies, we will get important clues on the beaming as well as the underlying true energy function.
Supervisor: Dr. Joscha Jahns-Schindler
-
Exploring the first proposed applications of Fast Radio Bursts
In this project, you will go on a historical journey about Fast Radio Bursts (FRBs). Only discovered in 2006, they are now used to tackle several astrophysical problems. However, long before their discovery, in 1965, they were already proposed as a probe for cosmology, namely to distinguish between a “de Sitter model” and a “steady state model” as the better description of our Universe. The distinction has long been done via other means, but it is interesting to try if the method could have worked and if today's data could distinguish between the two models. We will dive into old formulations of cosmological models and reformulate them in modern terms. Then we will use modern Bayesian methods to see if the proposed model distinctions can be done.
Further Reading:
Supervisor: Dr. Joscha Jahns-Schindler
-
Gravitational-wave data analysis with machine learning
Gravitational wave astronomy has offered an excitingly new perspective to astrophysics in the last few years. The field poses unique challenges in data analysis and instrumentation. Although sources of all the past observations have mainly been compact binary mergers, GW signals are expected to originate from a variety of other sources that can be observed by current ground-based detectors. Modelling of these sources must be done very accurately for their detection from the noisy data recorded by the detectors. The noise transients in the data also demand careful characterization to avoid false alarms. In this project, the student will study the basics of GW data analysis and subsequently develop the essential models for a unique set of GW sources, such as supernova explosions. The student will explore specialized deep learning techniques and / or statistical analysis in GW data analysis.
Further Reading:
Supervisor: Dr. Shreejit Jadhav
-
Detecting the first gravitational-wave signal from the explosion of massive stars
Hundreds of gravitational-wave signals have now been discovered from the merger of binary neutron stars and black holes, but other sources of gravitational waves have not yet been discovered. Some of the most violent explosive events in the Universe are predicted to emit bursts of gravitational waves, and may result in the next big multi-messenger discovery. One of the most promising astrophysical sources of gravitational waves is a core-collapse supernova. In this project, you will help to develop a new search for gravitational waves from core-collapse supernovae, and apply this new search to real data from the LIGO-Virgo-KAGRA gravitational-wave observatories.
Further Reading:
Supervisor: Dr. Jade Powell
-
The growing landscape of binary black hole formation scenarios
Gravitational waves from around 100 binary black hole mergers have been observed by Advanced LIGO and Virgo. The origin of binary black holes remains uncertain, with multiple proposed formation channels, including from isolated pairs of massive stars and dynamically assembled binaries in dense stellar environments. More recently, additional formation channels have been proposed, including from hierarchical stellar triples, from black holes trapped in disks of gas around active galactic nuclei, or even black holes of primordial origin. Each of these formation scenarios predicts unique, but overlapping distributions of binary parameters, such as mass, spin, eccentricity and redshift. In this project we will examine whether it is possible to constrain the relative contributions of each of these scenarios to the observed binary black hole population.
Further Reading:
Supervisor: Dr. Simon Stevenson
-
The formation of binary pulsars
Millisecond pulsars (MSPs) are believed to achieve their rapid rotation by accreting mass from a companion star in a close binary. Low-mass X-ray binaries (LMXBs) and Accreting Millisecond X-ray Pulsars (AMXPs) are observed during periods of active accretion, whilst so called ‘spider’ pulsars (‘redbacks’ and ‘black widows’) may represent a more evolved stage, where the companion is ablated by the intense radiation from the pulsar. In this project we will compute detailed binary stellar evolution models using the code MESA. In particular, we will examine the evolutionary stage prior to the LMXB stage accounting for recent advances in the understanding of magnetic braking of low-mass stars, and investigate the future evolution of such binaries (i.e., whether they evolve into redbacks or black widows).
Further Reading:
- MESA: Paxton et al. 2011
- Formation of Black Widows and Redbacks: Chen et al. 2013
- LMXBs: Van et al. 2018
Supervisor: Dr. Simon Stevenson