ISSC Videos
July 2, 2019
Daniel Muthukrishna, Cambridge University, speaks about his recent paper "RAPID: Early Classification of Explosive Transients using Deep Learning." Co-authors are Gautham Narayan, Kaisey S. Mandel, Rahul Biswas, and Renée Hložek. Paper is available at https://arxiv.org/abs/1904.00014. The presentation begins with comments by Chad Schafer on the general potential of deep learning in astronomy. Daniel's talk begins at 20:15.
April 29, 2020
Francois Lanusse, UC Berkeley, gives an introduction to neural network architectures useful for modelling time series.
April 29, 2020
Sara Jamal, UC Berkeley, presented results from her paper, "On Neural Network Architectures for Astronomical Time Series Classification," available at https://arxiv.org/abs/2003.08618. A recording of this presentation is available upon request.
May 8, 2020
Nic Dalmasso, Carnegie Mellon University, presents "Conditional Density Estimation Tools in Python and R". Slides are available here. Paper is available here (Arxiv version here). For the methods, the Jupyter notebooks tutorial are available on the respective Github repositories: NNKCDE RFCDE, FlexCode, DeepCDE. All tutorials also include the “cdetools” package for diagnostic, which is available here.
May 18, 2020
Ryan Hausen, UC Santa Cruz, speaks about his recent paper "Morpheus: A Deep Learning Framework For Pixel-Level Analysis of Astronomical Image Data" Co-author is Brant Robertson. Paper is available at https://arxiv.org/abs/1906.11248. More information regarding Morpheus, including tutorials, is available at https://morpheus-project.github.io/morpheus/.
July 7, 2020
Francois Lanusse speaks about the DESC Tomography Challenge. Slides available here and code and more information available at https://github.com/LSSTDESC/tomo_challenge
October 21, 2020
Anais Moller, CNRS, speaks on "Bayesian Neural Networks in Statistical Analyses: Towards Interpretability."
April 29, 2021
David Zhao, Carnegie Mellon University, presents "Validating Conditional Density Models and Posterior Estimates." The slides are available here. A paper, co-authored with Niccolo Dalmasso, Rafael Izbicki, and Ann Lee, describing this work is available on arXiv here, and software is available on GitHub here.