Striving to learn



  • 5 projects — [1] Prediction Neural Network, [2] CIFAR-10 Image Classification using Convolutional Neural Networks (CNN), [3] TV Script Generation using Recurrent Neural Network (RNN), [4] Language Translation using seq2seq model, [5] Face Generation using Generative Adversarial Networks(GANS)
    1. I used basic neural networks with the purpose of prediction of the number of bike requirements based on the data from the dataset, comparison of which activation function would best suit the dataset — sigmoid, heavy side, Rectified Linear Units, Softmax etc.; just an introductory project. Was introduced to TensorFlow and its amazing power coupled with simplicity. Also Worked on MNIST dataset with multilayer perceptron and performed basic sentimental analysis with TFLearn
    2. CIFAR-10 Image classifier, CNN’s were used to look, at any point in time at patches of images through a filter, whose size is defined by me. Multiple Neural Network connections were made to better classify image and capture all the required characteristics of an image.
    3. Used Long Short-Term Memory Networks (a part of RNN)for new TV script generation from existing data. Also used word2Vec architecture with the skip gram model to train a neural network with a single hidden layer, learned about experimenting and setting different hyper-parameters, which is necessary for any DL problem.
    4. Used seq2seq models with Tensorflow and transfer learning and reinforcement learning
    5. The most interesting project I’ve ever attempted, of creating new faces from a celebrity picture dataset using, fairly recent, Generative Adversarial Networks for generation of New face images and even MNIST digits.
  • Implemented Deep Neural Networks, Convolutional Neural Networks (CNNs), Recurrent Neural Networks(RNNs), Transfer Learning, Reinforcement Learning and Generative Adversarial Networks (GANS)
  • Worked on several Projects as a part of course work including a Language Translator, a CIFAR-10 Image Classifier, TV script Generation and Generation of faces using GANS
  • Language used for project: Python with Tensorflow(TF)


Check my certificate

Comments are closed.