Chris Pollett > Students >

    ( Print View)



    [CS297 Proposal]

    [Duolingo Activity List]

    [Feed-ForwardNN Slides-PDF]

    [Tensorflow Slides-PDF]

    [Convolution Neural Network Slides-PDF]

    [Word Embeddings And Recurrent NNS-PDF]

    [Sequence-to-Sequence Learning-PDF]

    [Deliverable 1]

    [Deliverable 2]

    [Deliverable 3]

    [Deliverable 4]

    [CS297 Final Report-PDF]

    [CS298 Proposal]

    [CS298 Final Report-PDF]

    [CS298 Slides-PDF]

Deliverable 1: Implement The Feed-Forward Neural Network using Tensorflow


Implement the forward feed neural network to recognize handwritten Gujarati language digits using Tensorflow. Handwritten Gujarati images are used as the train and test dataset. Images are scaled down to 16x16 pixels. Each digit image is converted to greyscale so each pixel can be categorized in either black or white color. Forward Feed Neural Network has different layers in its implementation. As shown in the image below, the entry layer feeds in image pixels to the perceptron. Each perceptron will assign a number to each possible outcome. Next layer is softmax layer, which provides a way to convert perceptron output into a probability distribution.

Source Code Download




[E. Charniak 2019] "Introduction to Deep Learning" . Eugene Charniak. The MIT Press. 2019.