About Schedule Presentations and Presenters Call for Presentations
- 1 Schedule
- 2 Presentations abstracts and presenter bios
- 2.1 Māori astronomy
- 2.2 Visualising the Open Universe
- 2.3 Open modelling of stars and galaxies from our own to those at the edge of the observable Universe
- 2.4 The SKA: Hacking the Big Bang
- 2.5 Period Analysis of lightcurves of MoA database: finding periods of variable stars
- 2.6 Data Mining the MOA Catalogue Using Machine-Learning Algorithms
- 2.7 GPU-accelerated Modeling of Microlensing Events
- 2.8 Lightning talks
Presentations abstracts and presenter bios
Dr. Pauline Harris
Dr Pauline Harris is from Rongomaiwahine and Ngāti Kahungunu. Pauline has completed a Masters Degree in Astronomy and has been awarded a PhD in Astrophysics from Canterbury University. Her research investigated Gamma Ray Bursts as possible sites for high-energy neutrino production. She is also a recipient of a Te Tipu Pῡtaiao Postdoctoral Award from the Foundation for Research, Science and Technology in the School of Chemical and Physical Sciences at Victoria University.
Her research is currently dedicated to searching for extra-solar planets and the collation and revitalisation of Māori Astronomical star lore. Pauline has been involved in the revitalisation of Māori star lore for the past 7 years and has given many talks both in New Zealand and overseas pertaining to Māori Astronomy and Matariki/Puanga.
Pauline is also heavily involved with Māori research in other areas as well, such as the relationship between Science and Mātauranga as well as within more specialised areas such as ethics in biotechnology. Dr Harris currently holds the position as Chair of the Society for Māori Astronomy, Research and Traditions (S.M.A.R.T) which has been established since 2009 and works to promote Astronomy and Natural Science as part of a Cultural and Scientific development incentive between Māori, Pacific and Polynesian communities and the wider New Zealand community. S.M.A.R.T aims to encourage the growth in numbers of Māori students, present and future, to pursue potential career paths within the Sciences, in particular, Astronomy. Dr Harris has co-authored two books and has published in a wide variety of forums.
Visualising the Open Universe
Astronomy is unparalleled in its power to captivate and educate. This is, in large part, owing to the availability of sensationally detailed and evocative imagery of celestial objects returned from high performance telescopes and other instrumentation. Much of that data is open domain. Similarly, some of the best simulation tools are available for free, including, for example, Stellarium. This talk highlights the opportunity for the open source community to contribute to developing tools specifically designed to use the available open data. An example of this would be to create data visualisation tools for VR hardware such as the Oculus Rift -- or the open hardware equivalent. Such tools can be pressed into use in teaching laboratories now using available data bases and should be in place before data from the next generation of large scale surveys come online, including the output from the Large Synopic Survey Telescope.
Dr. Nicholas James Rattenbury
Dr. Nicholas James Rattenbury is a Royal Society of New Zealand Rutherford Discovery Fellow. He completed his PhD in Physics at the University of Auckland and shortly thereafter left to do post-doctoral research at Jodrell Bank Observatory, The University of Manchester. After nearly five years of research, he worked for several years as a trainee patent attorney before returning to academia at Manchester Metropolitan University. As an RDF, he is returning to New Zealand to continue his research in astrophysics.
Open modelling of stars and galaxies from our own to those at the edge of the observable Universe
One way to understand the stars is to put all our understanding of physics into a computer code and create synthetic stars to predict how they evolve over their lifetimes and compare these to observations. The story begins during my PhD when computers had just started becoming fast enough that it was possible to create 100's of stellar models in a day. This for the first time enabled us to create sets of stellar models with different physics to see which was correct. This was done using a stellar evolution code originally written in the 70s that has always been freely available and updated/adapted by many researchers around the world. It is also the code that was adapted to make the current most open stellar evolution code MESA.
I adapted the code specifically to look at massive stars, those that end their lives in supernovae. Also I included the physics of interactions in a binary star system which at that point all other stellar models ignored. The problem with binary models is there are more free parameters so you need to calculate 1000's of stellar models to see the full diversity of evolution. Therefore I had to not only adapt a FORTRAN code I also had to put together a code to put organise and baby sit 1000s of models running at the same time. This was done in PERL.
Dr. Eldgridge will give an overview of the development of the BPASS (binary population and spectral synthesis) code which is the result of over a decade of work using open source tools and codes, and outline the next planed steps.
Dr. J.J. Eldgridge
Dr. JJ Eldridge is a Senior Lecturer of Astrophysics in the Department of Physics at the University of Auckland, and conducts general research is in the field of stellar evolution—particularly the effects of binary pathways on stellar populations, and the progenitors of supernovae. Since completing a PhD thesis at the Institute of Astronomy of Cambridge University, J.J. has worked as a postdoctoral researcher at Queen’s University Belfast and the Institut d’Astrophysique de Paris before returning to Cambridge in 2007 to work as a post-doc, and then commencing a lectureship in 2011.
The SKA: Hacking the Big Bang
The Square Kilometre Array (SKA) is an eleven countries mega-project -including New Zealand, to build a next generation radio telescope from 2018 till 2024 between Australia and South Africa. On an unprecedented scale, the SKA will revolutionise our understanding of the universe and the laws of fundamental physics; give insight into the formation of the first stars after the Big Bang; and, address one of humankind's ultimate questions: are we alone?
This fantastic facility will consist of thousands of telescopes distributed in several countries however the SKA will actually be the largest and most powerful high performance computer in the planet by far -with a radio telescope "attached" to it. The exascale computational requirements for the SKA are beyond the capabilities of existent technologies: they are needed to enable the SKA to scan the skies thousands of times faster than ever before, producing vast amounts of valuable data, at rates in to Tb/s (100x the global internet traffic).
The SKA's power will come from it being a software-controlled and software-dependent telescope. This will allow the SKA's designers to continuously take advantage of advances in computer power, algorithm design and data transport capability, always keeping SKA up to date. Around 500 engineers, scientists and researchers worldwide are contributing to the SKA’s design and development. Open Parallel (based in Oamaru, somewhere in the South Island of New Zealand) is leading the design of the Software Development Environment for the Central Signal Processor -the “brain” of the SKA.
In this talk, Nicolás will present the SKA and its incredible computing challenges; and share some thoughts on designing a software stack for massively distributed systems; and c) some questions -to share with the audience, about (other) possible applications of the SKA...
Nicolás Erdödy is the founder of Open Parallel, a company contributing to the Square Kilometre Array (SKA), the largest IT project in the world. He leads his own business consultancy and previously founded and managed a few ventures, including software, high-tech and e-learning companies. Born in Uruguay from Hungarian parents, he's now a kiwi in Oamaru where no one has a clue what he does for a living. His hobby is to collect citizenships and forgot FORTRAN decades ago but still knows how to ask for a beer in five or six human languages. He also organises Multicore World since 2012.
Period Analysis of lightcurves of MoA database: finding periods of variable stars
Microlensing of Astrophysics (MoA) is a world-leading collaboration between Japan and New Zealand on detecting exoplanets by microlensing techniques, which require long-term monitoring over the night sky with wide-field CCD camera to optimize the chance of any planet detections. As a result, numerous stars towards the Galactic bulge have been observed, night by night, over six years. Many of them should be variable stars which show either periodic or pseudo-periodic variability in their brightness.
In this talk, I will introduce the three period finding methods that I adopted in my research: conditional entropy, string-length method and phase dispersion method (PDM), compare their pros and cons with each other, and present the open-source codes of these methods that I write by modifying the exsiting open-source codes (e.g. PMD2 written by Stellingwerf) running on Linux.
Man Cheung Alex Li
Data Mining the MOA Catalogue Using Machine-Learning Algorithms
The Microlensing Observations in Astrophysics (MOA) project is a joint New Zealand-Japan astronomical collaboration, which has been running for the past two decades in order to search for gravitational microlensing events. Microlensing is a phenomenon predicted by Einstein’s general theory of relativity, and is characterised by the transient brightening of a background star due to the motion of an intermediate massive object near the line-of-sight. Observations made towards the galactic bulge by the MOA telescope—based in Canterbury—have detected thousands of microlensing events throughout the duration of the project. However, this characteristic signature is not unique to microlensing: many other astrophysical phenomena exhibit a similar temporal variability in brightness, and have also been catalogued. The result is that the MOA catalogue now contains millions of light curves—the vast majority of which have yet to be analysed— and is the largest astronomical database in New Zealand. The aim of my particular project is to make use of machine-learning (ML) algorithms in order to classify these light curves into distinct types of astrophysical events. To do so I am using “Weka”, an open-source software workbench of ML techniques developed by the Department of Computer Science at the University of Waikato. More specifically, I am using the Weka implementation of the random forest classifier, a tree-based supervised ML method. I propose to briefly outline the astrophysics of microlensing and the MOA project, before some discussion about ML in general; then a more in-depth analysis of the Weka implementation of the RF classifier and its application to this particular problem.
GPU-accelerated Modeling of Microlensing Events
Gravitational microlensing is a phenomenon whereby the light from a background source star is bent by the gravity of a foreground lens system causing the source star to appear to change its apparent brightness with time. These events are popularly being used to discover exoplanets. One of the techniques used to model these events is called inverse ray shooting which is a brute force method that shoots billions of rays from the observer plane through the lens plane onto the source plane. Inverse Ray Shooting is used to create magnification maps from which theoretical lightcurves can then be extracted for the purpose of model fitting to observational data.
GPU-accelerated inverse ray shooting accomplishes magnification map creation and source track extraction in seconds instead of hours as is the case with the CPU version of the same algorithm. This algorithm was developed by [cho_hong_ling_simulation_2013]and is run remotely on a desktop PC equipped with GPU archtiecture which is accessed via SSH capability on Linux.