Pursuit of Machine Learning

My experiments with ML

Dropout — A simple way to improve Neural Network ?

Dropout technique by Hinton is among recent pioneers in the field of neural network. The main objective behind using Dropout is to tackle the problem of overfitting. Dropout achieves this by selectively turning off (dropping out) some of the…

What is Adagrad ?

Adaptive Gradient Algorithm or Adagrad for short, is a general algorithm for choosing learning rate of a model by dynamically adapting to the dataset. It calculates different learning rate for every feature of model.

What Coursera’s Machine Learning course taught me?

Last December, I started taking Coursera’s Machine Learning course by Prof. Andrew Ng. This course has been a remarkable success since it’s inception in 2011. I would recommend everyone with interest in Machine Learning and related areas, to…