-
Prof. Richard McMahon18/09/2018, 10:00
Professor Richard McMahon, Professor of Astronomy, Director and Head of Department, Institute of Astronomy, University of Cambridge
Abstract: I will present an overview of non-radio imaging current and future challenges based on current ground and space based optical and near infrared imaging surveys; DES/VISTA/Gaia -> Euclid/LSST. Current surveys are producing PB scale imaging datasets at a...
Go to contribution page -
Prof. Stephen Smartt18/09/2018, 10:30
Prof. Stephen J. Smartt, Astrophysics Research Centre
School of Mathematics and Physics, Queen's University BelfastAbstract: Wide-field optical telescopes routinely employ large format detectors (CCDs) with 0.1 to 1 gigapixels which provide images every 30 seconds. Such facilities are capable of surveying the whole sky every night and the scientific exploitation requires immediate processing...
Go to contribution page -
Dr Bojan Nikolic18/09/2018, 10:50
Dr. Bojan Nikolic, Principal Research Associate at the Cavendish Laboratory, Cambridge
Abstract: I show that a software framework intended primarily for training of neural networks, PyTorch, is easily applied to a general function minimisation problem in science. The qualities of PyTorch of ease-of-use and very high efficiency are found to be applicable in this domain and lead to two...
Go to contribution page -
Prof. Jason McEwen (Astrophysics & Cosmology at Mullard Space Science Laboratory, University College London)18/09/2018, 11:10
Abstract: Over recent decades cosmology has transitioned from a data-poor to a data-rich field, which has lead to dramatic improvements in our understanding of the cosmic evolution of our Universe. Nevertheless, we remain ignorant of many aspects of the scenario that has been revealed. Little is known about the fundamental physics of structure formation in the early Universe or the formation...
Go to contribution page -
Dr Jennifer Thompson18/09/2018, 11:30
Dr. Jennifer Thompson, CERN
Abstract: The machine learning methods currently used in high energy particle physics often rely on Monte Carlo simulations of signal and background. A problem with this approach is that it is not always possible to distinguish whether the machine is learning physics or simply an artefact of the simulation. In this presentation I will explain how it is possible to...
Go to contribution page -
18/09/2018, 11:45
-
18/09/2018, 12:00
Moderator: Prof. Richard McMahon
Go to contribution page
Panelists: Dr. Shirley Ho -
Prof. Terry Lyons FLSW FRSE FRS18/09/2018, 14:00
Professor Terry Lyons FLSW FRSE FRS, Wallis Professor of Mathematics, Mathematical Institute, University of Oxford
Abstract: It often happens that measurements have redundant information over and above the invariants of interest; we might measure a rigid object in cartesian co-ordinates although we are only interested in the shape. Representing in an invariant way that does not carry...
Go to contribution page -
Dr Adam Elwood18/09/2018, 14:30
Dr. Adam Elwood
Abstract:"We introduce two new loss functions designed to directly optimise the statistical significance of the expected number of signal events when training neural networks to classify events as signal or background in the scenario of a search for new physics at a particle collider. The loss functions are designed to directly maximise commonly used estimates of the...
Go to contribution page -
Ms Vesna Lukic (PhD candidate University of Hamburg)18/09/2018, 14:45
Abstract: Machine learning techniques have proven to be increasingly useful in astronomical applications over the last few years, for example in object classification, estimating redshifts and data mining. A topic of current interest is to classify radio galaxy morphology, as it gives us insight into the nature of the AGN, surrounding environment and evolution of the host galaxy. The task of...
Go to contribution page -
Dr Hao Ni (Senior Lecturer in Financial Mathematics at UCL and the Turing Fellow at The Alan Turing Institute )18/09/2018, 15:00
Abstract: In this talk, we consider the supervised learning problem where the explanatory variable is a data stream. We provide an approach based on identifying carefully chosen features of the stream which allows linear regression to be used to characterise the functional relationship between explanatory variables and the conditional distribution of the response; the methods used to develop...
Go to contribution page -
Dr Andrey Ustyuzhanin18/09/2018, 15:30
Dr. Andrey Ustyuzhanin, Head of Yandex-CERN Joint Research Projects & Head of the Laboratory of Methods for Big Data Analysis at National Research University Higher School of Economics
Abstract: There is an exceptional way of doing data-driven research employing networked community. The following examples can illustrate the approach: Galaxy Zoo or Tim Gower’s blog. However many cases of...
Go to contribution page -
Dr Griffin Foster (University of Oxford, University of California at Berkeley)18/09/2018, 16:00
Abstract: Breakthrough Listen is the largest effort to date to search for techno-signatures from extraterrestrial civilizations. We use extensive computing power to search at high frequency and time resolution for transients events in petabytes of observational data from the Green Bank Telescope, Parkes Telescope, LOFAR, and soon MeerKAT. Given the diverse manifestations of transient signals...
Go to contribution page -
Dr Robert Lyon18/09/2018, 16:15
Abstract: To harness the discovery potential of data collected by the SKA, we require efficient and effective automated data processing methods. Machine learning tools have the potential to deliver this capability, as evidence via their successful application to similar problems in the astronomy domain. This talk introduces the machine learning required for successful time-domain data...
Go to contribution page -
18/09/2018, 17:00
Moderator: Prof. Paul Alexander
Go to contribution page
Panelists: Dr. Peter Braam, Prof. Terry Lyons, Prof. Richard McMahon, Prof. Ian Shipsey -
Choose timezone
Your profile timezone: