Long Short-Term Memory (LSTM)

YouTube Search ...Google search

• Neural Network Zoo | Fjodor Van Veen
• A Beginner's Guide to LSTMs
• Recurrent Neural Network (RNN) Variants:
  • Long Short-Term Memory (LSTM)
  • Gated Recurrent Unit (GRU)
  • Bidirectional Long Short-Term Memory (BI-LSTM)
  • Bidirectional Long Short-Term Memory (BI-LSTM) with Attention Mechanism
  • Average-Stochastic Gradient Descent (SGD) Weight-Dropped LSTM (AWD-LSTM)
• Andrej Karpathy blog
• Understanding LSTM Networks | Christopher Olah
• Animated RNN, LSTM and GRU | Raimi Karim - Towards Data Science
• Predicting Stock Price with LSTM | Asutosh Nayak - Towards Data Science ...Code
  • Stock Buy/Sell Prediction Using Convolutional Neural Network | Asutosh Nayak - Towards Data Science

A LSTM (Long Short-term Memory) Neural Network is just another kind of Artificial Neural Network, which falls in the category of Recurrent Neural Network (RNN). What makes LSTM Neural Networks different from regular Neural Networks is, they have LSTM cells as neurons in some of their layers. Much like Convolutional Layers help a (Deep) Convolutional Neural Network (DCNN/CNN) learn about image features, LSTM cells help the Network learn about temporal data, something which other Machine Learning models traditionally struggled with. ... Each LSTM cell in our Neural Network will only look at a single column of its inputs, and also at the previous column’s LSTM cell’s output. Normally, we feed our LSTM Neural Network a whole matrix as its input, where each column corresponds to something that “comes before” the next column. This way, each LSTM cell will have two different input vectors: the previous LSTM cell’s output (which gives it some information about the previous input column) and its own input column. LSTM: How To Train Neural Networks To Write Like Lovecraft | Strikingloo