Final Exam: DL Programmer

WHAT YOU WILL LEARN

• 

  Distinguish been input, output, and hidden layers in a neural network
  

  recognize the different types of neural network computational models
  

  describe resnet layers and blocks
  

  implement long short-term memory using tensorflow
  

  compare the supervised and unsupervised learning methods of artificial neural networks
  

  list neural network algorithms that can be used to solve complex problems across domains
  

  recall the essential hyperparameters that are applied on convolutional networks for optimization and model refinement
  

  recall the approaches of identifying overfitting scenarios and preventing overfitting using regularization techniques
  

  describe gradient descent and list its prominent variants
  

  recognize the need for activation layer in convolutional neural networks and compare the prominent activation functions for deep neural networks
  

  identify the need for activation layer in convolutional neural networks and compare the prominent activation functions for deep neural networks
  

  define and classify activation functions and provide a comparative analysis with the pros and cons of the different types of activation functions
  

  recognize the machine learning problems that we can address using hyperparameters along with the various hyperparameter tuning methods and the problems associated with hyperparameter optimization
  

  define multilayer perceptrons and illustrate the algorithmic difference from single layer perceptrons
  

  describe functions in calculus
  

  demonstrate how to test multiple models and select the right model using scikit-learn
  

  describe the approach of creating deep learning network models along with the steps involved in optimizing the networks
  

  describe shared parameters and spatial in a convolutional neural network (cnn)
  

  demonstrate how to build a convolutional neural network for image classification using python
  

  describe the concept of scaling data and list the prominent data scaling methods
  

  define semantic segmentation and its implementation using texton forest and random-based classifier
  

  define the concepts of variance, covariance and random vectors
  

  list the essential clustering techniques that can be applied on artificial neural network
  

  use backpropagation and keras to implement multi-layer perceptron or neural net
  

  identify and illustrate the use of learning rates to optimize deep learning
  

  build a recurrent neural network using pytorch and google colab
  

  describe vanishing gradient problem implementation approaches
  

  describe the regularization techniques used in deep neural network
  

  develop convolutional neural network models from the scratch for object photo classification using python and keras
  

  build neural networks using pytorch
• 

  specify approaches that can be used to implement predictions with neural networks
  

  recognize the need for gradient optimization in neural networks
  

  implement convolutional neural networks (cnns) using pytorch
  

  work with hyperparameters using keras and tensorflow to derive optimized convolutional network models
  

  demonstrate how to select hyperparameters and tune for dense networks using hyperas
  

  define the concept of the edge detection method and list the common algorithms that are used for edge detection
  

  recognize the various approaches of improving the performance of machine learning using data, algorithm, algorithm tuning and ensembles
  

  identify the different types of learning rules that can be applied in neural networks
  

  build deep learning language models using keras
  

  describe the purpose of a training function in an artificial neural network
  

  calculate loss function and score using python
  

  recognize the limitations of sigmoid and tanh and describe how they can be resolved using relu along with the significant benefits afforded by relu when applied in convolutional networks
  

  implement backpropagation using python to train artificial neural networks
  

  recognize the differences between the non-linear activation functions
  

  recognize the role of pooling layer in convolutional networks along with the various operations and functions that we can apply on the layer
  

  describe the temporal and heterogeneous approaches of optimizing predictions
  

  implement recurrent neural network using python and tensorflow
  

  list activation mechanisms used in the implementation of neural networks
  

  describe sequence modeling as it pertains to language models
  

  implement calculus, derivatives, and integrals using python
  

  implement the artificial neural network training process using python
  

  list features and characteristics of gated recurrent units (grus)
  

  demonstrate the implementation of differentiation and integration in r
  

  work with threshold functions in neural networks
  

  describe the iterative workflow for machine learning problems with focus on essential measures and evaluation protocols
  

  recognize the importance of linear algebra in machine learning
  

  define and illustrate the use of learning rates to optimize deep learning
  

  recall the prominent optimizer algorithms along with their properties that can be applied for optimization
  

  recognize the involvement of maths in convolutional neural networks and recall the essential rules that are applied on filters and channel detection
  

  recall the algorithms that can be used to train neural networks