Dananjaya Liyanage's Projects
In this repo we will develop new transfer learning emulation methods that can be used when source and target have different model parameters. We will use Trento-2D and Trento-3D models with pseudo observables as our test cases for new TL methods.
Efficient emulation with transfer learning techniques.
Heavy-ion collision initial condition model
How much do you trust the predictions from your machine learning model? Depending on where you apply the machine learning model this question might be a matter of life and death. Can we identify the possible sources that contribute to make your machine learning model predictions inaccurate? If we can not completely get rid of these uncertainties can we at least quantify it? What are the ways that we can quantify the uncertainty in machine learning model predictions? These are the questions that we will try to find answers in "Uncertainty in ML" working group in Erdos boot camp in the fall of 2021.
This is a streamlit widget made to visualize the initial simulation data from the VAH project. This is largely based on a widget developed by Derek Everett.