wayground logo

Free Printable Worksheets

Font size

S
M
L
XL
Worksheets

Understanding Recurrent Neural Networks

Total questions: 11

Worksheet time: 6mins

Name
Class
Date
1.

What is the purpose of concatenating the outputs of the forward and backward RNNs?

a)

To form a combined representation

b)

To discard unnecessary information

c)

To simplify the model

d)

To increase computational cost

2.

What are some applications of Bi-RNNs?

a)

Sentiment Analysis

b)

Image Classification

c)

Data Compression

d)

Time Series Forecasting

3.

What is one advantage of Bidirectional RNNs?

a)

Enhanced Contextual Understanding

b)

Lower computational cost

c)

Simpler architecture

d)

Faster training times

4.

What is a drawback of Bidirectional RNNs?

a)

Increased Computational Complexity

b)

Better accuracy

c)

More efficient memory usage

d)

Faster predictions

5.

What types of Bi-RNNs are mentioned?

a)

LSTM Bi-RNNs

b)

GRU Bi-RNNs

c)

CNN Bi-RNNs

d)

SVM Bi-RNNs

6.

What metrics are used to evaluate traditional language modeling?

a)

Accuracy

b)

F1-Score

c)

Perplexity

d)

Cross entropy

7.

What is the perplexity metric in NLP?

a)

A way to capture the degree of uncertainty a model has in predicting text.

b)

A measure of the average uncertainty in predicting words.

c)

A metric used to evaluate the performance of language models.

8.

How is perplexity computed over a sentence W?

a)

PP(W) = 1/Pnorm(W)

b)

PP(W) = P(W)^(1/n)

c)

PP(W) = Pnorm(W)^(1/n)

9.

What does cross-entropy measure?

a)

How well the predicted distribution aligns with the true distribution.

b)

The average uncertainty in predicting words.

c)

The performance of language models.

10.

What is Bits-per-Character (BPC)?

a)

A metric used to evaluate the performance of language models.

b)

A measure of the average number of bits required to represent a character in a given text.

c)

A method to compute the perplexity of a language model.

11.

What are the problems with RNNs?

a)

They are unable to make predictions based on past, present, and future events simultaneously.

b)

They struggle with tasks requiring deep contextual understanding.

c)

They have fixed input size and no time dependencies.