Ashutosh singh , E-mail: ashutoshsingh9798.as@gmail.com
Hi, I am Ashutosh singh and i would like to share this project because it just shows how a machine can learn and perform a task the same way we Humans do. It is just a simple project of Hand Written Digit Recognition system , However Machine can do more complex task such as Object Detection, Facial recognition, Speech recognition and so on.Humans need Year of practice and observation to do these type of things but through Machine Learning we can tech a machine to do it in just few minutes. Let us take an example of yourself , You just take years to recognise faces, speeches and objects in your childhood and currently you are also learning the same way. You just learn day after day and making your brain more and more powerful. Machines can also work in the simmilar manner the only difference is that We have inbuilt sensors in our body(like eyes,nose, skin, and various others) to get the data from the universe by our own but we havee to feed the data manually for machines or to build a sensor like camera , microphone e.t.c in it so that it can collect data. Once it has collected data it doesn't take that much of time to learn as we do. Nowadays We are trying to surpass the Human level of Intelligence by these machines power what we called "Artificial Intelligence".
This is a part of a Assignment in Andrew Ng's M.L Course. I have Completed this course a couple of months ago with a grade of 96.1% In this Project we have taken Our data from MNIST HAnd written digit recognition and train it in a Deep Neural Network. We have got Training Accuracy Of 99%. All the computation of different algorithm are vectorised and have tried to avoid "for loops" as much as possible so that our code runs faster. I have not Explained the Working of The Backpropagation Algorithm and from where that vectorised implementation come from and have just written the code if You want the explanation you can mail me. For More Information take a look on the Pdf provided, all other informations are mentioned there.