lstm_text

Text Classification with Long Short-Term Memory (LSTM) model
Last Queried: Nov. 28, 2019, 6:26 p.m.

Producer: ahmadhusain

Star Follow @ahmadhusain

  • You can participate in this project.

    Details

lstm_text Download

No votes
Visit on GitHub
Preview Quiz.md
  1. When building an architecture neural network with a large number of hidden layer. We may face a problem, the update value from backpropogation steps will exponentially decrease when heading to the input layer. So the model fails to learn when doing a pass forward to make a predictions. Which of the following statement is appropriate to that problem?
    • [ ] Exploding Gradient Descent
    • [ ] Vanishing Gradient Descent
    • [ ] Backpropogarion Through Time
  2. Long Short-Term memory (LSTM) architecture using gate mechanism to overcome a gradient descent's problem. Which gate that aims to controls how many internal states (information) want to exposes to the next time step?
    • [ ] Output gate
    • [ ] Hidden gate
    • [ ] Input gate
  3. When working with text dataset on classification task, there is tokenization step that aims to separate each word in the entire document into a token form. We need to do feature selection before modeling. Which of the argument on text_tokenizer below is appropriate to setting the maximum number of words to be used?
    • [ ] lower
    • [ ] num_words
    • [ ] fit_text_tokenizer
  4. If we want to create matrix result of the transformation each text in a sequence of integer, which of the function below is appropriate?
    • [ ] text_tokenizer()
    • [ ] pad_sequences()
    • [ ] text_to_sequences()
  5. When building an architecture model, we want a model does not "memorize" the training data, but learns the actual relationship on data! Which of the following methods can we use to avoid underfitting and overfitting?
    • [ ] added layer_dropout with dropout rate closer to 1.
    • [ ] added layer_dropout with dropout rate between 0.2 to 0.5
    • [ ] increase number of epoch when train the model.
  6. Suppose we want to predict unseen text data, of course we need to do pre-processing step first. Which of information from model that we need to keep when doing text-processing?
    • [ ] tokenizer
    • [ ] epoch
    • [ ] batch size
Quiz
You need to score 4 out of a possible 6 to earn a badge.
You have 2 attempts. Only your highest score will be taken into account.
  • Quiz 1

  • When building an architecture neural network with a large number of hidden layer. We may face a problem, the update value from backpropogation steps will exponentially decrease when heading to the input layer. So the model fails to learn when doing a pass forward to make a predictions. Which of the following statement is appropriate to that problem?

    • Question worth 1 point
  • Quiz 2

  • Long Short-Term memory (LSTM) architecture using gate mechanism to overcome a gradient descent's problem. Which gate that aims to controls how many internal states (information) want to exposes to the next time step?

    • Question worth 1 point
  • Quiz 3

  • When working with text dataset on classification task, there is tokenization step that aims to separate each word in the entire document into a token form. We need to do feature selection before modeling. Which of the argument on text_tokenizer below is appropriate to setting the maximum number of words to be used?

    • Question worth 1 point
  • Quiz 4

  • If we want to create matrix result of the transformation each text in a sequence of integer, which of the function below is appropriate?

    • Question worth 1 point
  • Quiz 5

  • When building an architecture model, we want a model does not "memorize" the training data, but learns the actual relationship on data! Which of the following methods can we use to avoid underfitting and overfitting?

    • Question worth 1 point
  • Quiz 6

  • Suppose we want to predict unseen text data, of course we need to do pre-processing step first. Which of information from model that we need to keep when doing text-processing?

    • Question worth 1 point
Recipients 0

Users with passing score on lstm_text

    No recipients yet. Be the first!