AI-ML Projects

Computed sampling distributions of mean and variance, and calculated confidence intervals

Hypothesized that solar eclipses averagely are visible from the equator tested it with Z and T tests

Performed regression analysis on the variables and fitted a linear relation with Least Squares Estimation

Scrutinized research literature on CNN architectures for emotion recognition using human face data

Incorporated dropouts and L2 reg. for generalization; Enabled batch training for efficiency

Applied Stochastic Gradient Descent & Binary Cross Entropy loss and achieved a test-set accuracy of 60%