Name: Aleksandra Ciprijanovic
Type: User
Company: Fermi National Accelerator Laboratory
Bio: Astrophysicist working at the intersection of astronomy and machine learning. True believer in ML for science.
Twitter: ACiprijanovic
Location: Batavia
Blog: https://aleksandraciprijanovic.wordpress.com/
Aleksandra Ciprijanovic's Projects
My Website
AstroHackWeek2020 Hack proposal #Astro-MNIST
Basic alrothms for classification, clustering and Neural Networks (me practicing stuff with sentdex).
Intro to Python / data science curriculum for Berkeley Lab Director's Apprenticeship Program
CNN for differentiating between cats and dogs in TensorFlow.
Data and scripts used for the cosmology with one galaxy paper
Domain Adversarial Neural Network approach to galaxy merger classification using pristine and noisy images together
Studying robustness of deep learning algorithms to adversarial attacks that may naturally arise from scientific data pipelines, such as LSST.
Using Contrastive Domain Discrepancy for class-aware domain adaptation to study galaxy mergers.
DeepMerge: An implementation of neural network models for classifying merging galaxies
Fisher deep domain adaptation for astronomy - galaxy mergers
Repository for the project "DeepShadows: Separating LSBGs from artifacts using Deep Learning"
demo
A validation framework and tests for mock galaxy catalogs and beyond
code for "Fisher Deep Domain Adaptation"
Astro Hack Week Hacks Collector: a repository for past and present dotAstronomy hacks
A toolkit to get you started with organizing your own hack week!
Uses neural networks to extract sufficient statistics from data by maximising the Fisher information
Mask R-CNN for object detection and instance segmentation on Keras and TensorFlow
PyTorch experiments
Random usefull programs
Tutorial Jupyter Notebooks for Data Preview 0, created and maintained by the Rubin Observatory Community Engagement Team.
Approximating Wasserstein distances with PyTorch