The goal is to employ a systematic framework grounded in probabilistic reasoning and optimization, in order to gain a fundamental understanding of the “best” approaches out there for topics such as unsupervised and supervised learning, sparsity-centric techniques, and Monte Carlo techniques.
from root folder:
python -m path.to.module
-
Model_and_data_based_classification/map_rule.py: Compare model-based (MAP) and data-based (logistic regression) classification on generated 2D-gaussian-mixture samples.
Link to report -
Neural_Network_classification/fully_connected_NN.py: Classification of synthetic 2D-gaussian-mixture samples with neural networks.
Link to report -
Unsupervised_Learning/unsupervised_learning.py: Compare unsupervised learning methods such as KMeans, Gaussian mixtures and the EM algorithms.
Link to report -
PCA_and_Compressive_Sensing/sparisity.py: Study sparsity (PCA and Compressive Sensing).
Link to report