Neural-Networks-and-Deep-Learning

Shristi Kedia(BITS Pilani)

• This is my assignment on Andrew Ng's "Deep Learning Specialization"; a 5 course specialization consisting of the following:

  • Neural Networks and Deep Learning
  • Improving Deep Neural Networks: Hyperparameter tuning, Regularization and Optimization
  • Structuring Machine Learning Projects
  • Convolutional Neural Networks
  • Sequence Models

Course 1 : Neural Networks and Deep Learning

• In this course, the foundations of deep learning were covered. The major learnings after completing this course were :
  • Understood the major technology trends driving Deep Learning.
  • Been able to build, train and apply Fully Connected Deep Neural Networks.
  • Leant how to implement efficient (Vectorized) neural networks.
  • Understood the key parameters in a neural network's architecture.
  • This course also taught how Deep Learning actually works, rather than presenting only a cursory or surface-level description.

Programming Assignments :

• The assignments of this course are in the folder Course 1 and it consits of the following assignments:
  • Python Basics with Numpy : [notebook]
  • Logistic Regression with a Neural Network mindset : [notebook]
  • Planar data classification with one hidden layer : [notebook]
  • Building your Deep Neural Network Step by Step : [notebook]
  • Deep Neural Network Application : [notebook]

Course 2 : Improving Deep Neural Networks: Hyperparameter tuning, Regularization and Optimization

• After finishing the course the major takeaways were:
  • Understood industry best-practices for building deep learning applications.
  • Been able to effectively use the common neural network "tricks", including Initialization, L2 and Dropout Regularization, Batch Normalization, Gradient Checking.
  • Been able to implement and apply a variety of optimization algorithms, such as mini-batch Gradient Descent, Momentum, RMSprop and Adam, and check for their convergence.
  • Understand new best-practices for the deep learning era of how to set up train/dev/test sets and analyze Bias/Variance.
  • Been able to implement a neural network in TensorFlow.

Programming Assignments :

• The assignments of this course are in the folder Course 2 and it consits of the following assignments:
  • Initialization : [notebook]
  • Regularization : [notebook]
  • Gradient Checking : [notebook]
  • Optimization : [notebook]
  • Tensorflow Tutorial : [notebook]

Course 3 : Structuring Machine Learning Projects

• This course taught how to build a successful machine learning project. The major learning after finishing this course were as follows :
  • Understood how to Diagnose Errors in a machine learning system, and been able to prioritize the most promising directions for reducing error.
  • Understood complex ML settings, such as Mismatched training/test sets, and comparing to and/or surpassing human-level performance
  • Learnt how to apply End-To-End Learning, Transfer Learning, and Multi-Task Learning.

Course 4 : Convolutional Neural Networks

• This course taught how to build Convolutional Neural Networks and apply it to image data. The major learnings after completing this course were :
  • Understand how to build a convolutional neural network, including recent variations such as residual networks.
  • Know how to apply convolutional networks to visual detection and recognition tasks.
  • Know to use neural style transfer to generate art.
  • Be able to apply these algorithms to a variety of image, video, and other 2D or 3D data.

Programming Assignments :

• The assignments of this course are in the folder Course 4 and it consits of the following assignments:
  • Convolutional Model:step by step : [notebook]
  • Convolution model Application : [notebook]
  • Keras Tutorial Happy House : [notebook]
  • Residual Networks : [notebook]
  • Autonomous driving application Car detection : [notebook]
  • Art Generation with Neural Style Transfer : [notebook]
  • Face Recognition for the Happy House : [notebook]

Course 5 : Sequence Models

• This course taught how to build models for Natural Language, Audio, and other Sequence Data. After finishing the course the major takeaways were :
  • Understood how to build and train Recurrent Neural Networks (RNNs), and commonly-used variants such as GRUs and LSTMs.
  • Been able to apply sequence models to natural language problems, including Text Synthesis.
  • Been able to apply sequence models to audio applications, including Speech Recognition and Music Synthesis.

Programming Assignments :

• The assignments of this course are in the folder Course 5 and it consits of the following assignments:
  • Building a Recurrent Neural Network Step by Step : [notebook]
  • Dinosaurus Island Character level language model : [notebook]
  • Jazz improvisation with LSTM : [notebook]
  • Operations on word vectors : [notebook]
  • Emojify : [notebook]
  • Neural machine translation with attention : [notebook]
  • Trigger word detection : [notebook]