VanillaRNN and LSTM

1234567891011
Across
  1. 3. Method for training neural networks
  2. 5. Refers to the problem of learning from information that is many steps back in time
  3. 7. Part of LSTM that decides what information to output from the cell state
  4. 8. Refers to the slope of the neural network's loss function
  5. 11. Information transferred from one step to the next in RNNs and LSTMs
Down
  1. 1. Part of LSTM that decides what information to discard from the cell state
  2. 2. Component of LSTM that stores the state information
  3. 4. Basic type of recurrent neural network
  4. 6. Part of LSTM that decides what new information to store in the cell state
  5. 9. Neural network architecture that avoids the vanishing gradient problem
  6. 10. Mechanism used by LSTM to control the flow of information