Neural Networks applied to Data Analysis in Cosmology
Neural Networks applied to Data Analysis in Cosmology
In the year 2000, I've published a paper where a neural network was applied to simulated observations of the Cosmic Microwave Background. The algorithm was capable of learning the statistical properties of processed signals linearly mixed together and separating them out, similar to the "cocktail party problem". The inverse linear filter, called W, was updated at each step (each recorded sky signal) through the simple equation
In the year 2000, I've published a paper where a neural network was applied to simulated observations of the Cosmic Microwave Background. The algorithm was capable of learning the statistical properties of processed signals linearly mixed together and separating them out, similar to the "cocktail party problem". The inverse linear filter, called W, was updated at each step (each recorded sky signal) through the simple equation
