Network training - Examples

Boolean

Dummy dataset

• Basics with Perceptron (P)
  • Import, configure and retrieve data
    • Imports
    • Configuration
    • Retrieve Boolean features and label
  • Perceptron - Single layer Neural Network
    • The Embedding layer object
    • The dataSet object
    • The Dense layer object
  • The EpyNN Network object
    • Instantiate your Perceptron
    • Perceptron training
  • Write, read & Predict

String

Dummy dataset

• Basics with string sequence
  • Environment and data
  • Feed-Forward (FF)
    • Embedding
    • Flatten-Dense - Perceptron
    • Difference between accuracy and cost
  • Recurrent Architectures
    • Embedding
    • RNN-Dense
    • LSTM-Dense
    • GRU-Dense
  • Write, read & Predict

Protein Modification

• O-GlcNAc Prediction
  • Environment and data
  • Feed-Forward (FF)
    • Embedding
    • Flatten-(Dense)n with Dropout
  • Recurrent Architectures
    • Embedding
    • LSTM-Dense
    • LSTM(sequence=True)-Flatten-Dense
    • LSTM(sequence=True)-Flatten-(Dense)n with Dropout
  • Write, read & Predict

Numerical (time-series)

Dummy dataset

• Basics with numerical time-series
  • Environment and data
  • Feed-Forward (FF)
    • Embedding
    • Flatten-(Dense)n
    • Flatten-(Dense)n with Dropout
  • Recurrent Neural Network (RNN)
    • Embedding
    • RNN-Dense
    • RNN-Dense with SGD
  • Write, read & Predict

Music Authorship

• Distinguish author-specific patterns in music
  • Environment and data
  • Feed-Forward (FF)
    • Embedding
    • Flatten-(Dense)n with Dropout
  • Recurrent Architectures
    • Embedding
    • RNN(sequences=True)-Flatten-(Dense)n with Dropout
    • GRU(sequences=True)-Flatten-(Dense)n with Dropout
  • Write, read & Predict

Numerical (images)

Dummy dataset

• Basics with images
  • Environment and data
  • Feed-Forward (FF)
    • Embedding
    • Flatten-(Dense)n with Dropout
  • Convolutional Neural Network (CNN)
    • Embedding
    • Conv-Pool-Flatten-Dense
  • Write, read & Predict

MNIST Database

• Multiclass Classification
  • Environment and data
  • Feed-Forward (FF)
    • Embedding
    • Flatten-(Dense)n with Dropout
  • Convolutional Neural Network (CNN)
    • Embedding
    • Conv-MaxPool-Flatten-Dense
    • Conv-MaxPool-Flatten-(Dense)n
  • Write, read & Predict