Deep Learning Questions | Practice Test / Quiz / MCQs

Home > Machine Learning 101 > Quizzes > Deep Learning Questions

Deep Learning Questions

Fast practice, instant feedback. Timer auto-submits when time’s up.

Avg score: 40% • Most missed: “What is an NVP network?”

Study First Challenge a Friend Stats

Deep Learning Questions

Time left 00:00

25 Questions

1. Are word embeddings trained on supervised" or "un-supervised" data?"

Such a model can use an encoder-decoder. For example, a translator which translates french to english; if the sentence in french has 10 words, that does not mean that the sentence in english will also have 10 words.It is where one network tries to spot the mistakes of another one, while this network tries to fool the spotter" network."Un-supervised. This is a big benefit as it allows to train on potentially huge corpora of data.True. During inference, batch normalization shifts and rescales independently each component of the input x according to statistics estimated during training.

2. True/False. During inference, batch normalization performs a component-wise aï¬ƒne transformation.

Such a model can use an encoder-decoder. For example, a translator which translates french to english; if the sentence in french has 10 words, that does not mean that the sentence in english will also have 10 words.It is where one network tries to spot the mistakes of another one, while this network tries to fool the spotter" network."True. During inference, batch normalization shifts and rescales independently each component of the input x according to statistics estimated during training.Un-supervised. This is a big benefit as it allows to train on potentially huge corpora of data.

3. How can a sequence-2-sequence model have a differently sized inputs and outputs?

It is where one network tries to spot the mistakes of another one, while this network tries to fool the spotter" network."Un-supervised. This is a big benefit as it allows to train on potentially huge corpora of data.True. During inference, batch normalization shifts and rescales independently each component of the input x according to statistics estimated during training.Such a model can use an encoder-decoder. For example, a translator which translates french to english; if the sentence in french has 10 words, that does not mean that the sentence in english will also have 10 words.

4. What are adversarial networks?

Such a model can use an encoder-decoder. For example, a translator which translates french to english; if the sentence in french has 10 words, that does not mean that the sentence in english will also have 10 words.Un-supervised. This is a big benefit as it allows to train on potentially huge corpora of data.True. During inference, batch normalization shifts and rescales independently each component of the input x according to statistics estimated during training.It is where one network tries to spot the mistakes of another one, while this network tries to fool the spotter" network."

5. What is the Wasserstein distance?

Intuitively, it is the minimum mass displacement to transform one distribution into the other. It solves the problem of the JS divergence regarding the structure of the distributions.True. During inference, batch normalization shifts and rescales independently each component of the input x according to statistics estimated during training.Such a model can use an encoder-decoder. For example, a translator which translates french to english; if the sentence in french has 10 words, that does not mean that the sentence in english will also have 10 words.Un-supervised. This is a big benefit as it allows to train on potentially huge corpora of data.

6. What does the Skip Gram model do?

Un-supervised. This is a big benefit as it allows to train on potentially huge corpora of data.Such a model can use an encoder-decoder. For example, a translator which translates french to english; if the sentence in french has 10 words, that does not mean that the sentence in english will also have 10 words.True. During inference, batch normalization shifts and rescales independently each component of the input x according to statistics estimated during training.The skip gram model is an algorithm which optimizes word embeddings so that a word can be predicted by a single word in its context. It is the opposite of CBOW, where a context is used to predict a word. Here, a word is essentially used to predict a context.

7. What is the general idea behind a convolution layer?

True. During inference, batch normalization shifts and rescales independently each component of the input x according to statistics estimated during training.Un-supervised. This is a big benefit as it allows to train on potentially huge corpora of data.A representation meaningful at a certain location can / should be used everywhere. It applies the same linear transformation locally, everywhere, and preserves the signal structure.Such a model can use an encoder-decoder. For example, a translator which translates french to english; if the sentence in french has 10 words, that does not mean that the sentence in english will also have 10 words.

8. What is the universal approximation theorem?

Any continuous function can be approximated with a single layer perceptron. IE, a combination of ReLUs can approximate any continuous function. It states that the training error can be as low as we want by making the layer larger. However, nothing is stated about the test error.Such a model can use an encoder-decoder. For example, a translator which translates french to english; if the sentence in french has 10 words, that does not mean that the sentence in english will also have 10 words.True. During inference, batch normalization shifts and rescales independently each component of the input x according to statistics estimated during training.Un-supervised. This is a big benefit as it allows to train on potentially huge corpora of data.

9. Is the empirical risk a biased or an unbiased estimator of the risk?

True. During inference, batch normalization shifts and rescales independently each component of the input x according to statistics estimated during training.It is an unbiased estimator of the risk.Un-supervised. This is a big benefit as it allows to train on potentially huge corpora of data.Such a model can use an encoder-decoder. For example, a translator which translates french to english; if the sentence in french has 10 words, that does not mean that the sentence in english will also have 10 words.

10. What is Spectral Normalization and why is it implemented in GANs?

True. During inference, batch normalization shifts and rescales independently each component of the input x according to statistics estimated during training.Spectral Normalization is a layer normalization that estimates the largest singular value of a weight matrix, and re-scale it accordingly. It was proposed to control the Lipschitz constant of D by rescaling its weights, so that all the linear layers have their singular values lesser than 1, and consequently Lipschitz constant, lesser than 1. If the non-linear layers are also Lipschitz of constant lesser than 1 (e.g. ReLU), this is a suï¬ƒcient condition.Un-supervised. This is a big benefit as it allows to train on potentially huge corpora of data.Such a model can use an encoder-decoder. For example, a translator which translates french to english; if the sentence in french has 10 words, that does not mean that the sentence in english will also have 10 words.

11. What is the concatenated ReLU?

x --> (max(0,x),max(0,âˆ’x)), it doubles the amount of activations, but maintains the norm of the signal intact during both the forward and backward passes.Un-supervised. This is a big benefit as it allows to train on potentially huge corpora of data.True. During inference, batch normalization shifts and rescales independently each component of the input x according to statistics estimated during training.Such a model can use an encoder-decoder. For example, a translator which translates french to english; if the sentence in french has 10 words, that does not mean that the sentence in english will also have 10 words.

12. Why is gating used in RNN?

When unfolding through time, the models get deep and can run into the vanishing gradient problem. TO solve this, pass-through" gates allow the recurrent state to avoid going repeatedly through squashing non-linearities."True. During inference, batch normalization shifts and rescales independently each component of the input x according to statistics estimated during training.Un-supervised. This is a big benefit as it allows to train on potentially huge corpora of data.Such a model can use an encoder-decoder. For example, a translator which translates french to english; if the sentence in french has 10 words, that does not mean that the sentence in english will also have 10 words.

13. What is the main weakness of the original for seq2seq translation? This weakness is fixed by attention mechanisms.

Un-supervised. This is a big benefit as it allows to train on potentially huge corpora of data.All the information must flow through a single state v, which must have the capacity for any situation. Additionally, there are no direct channels to transport local information to where it is useful. Attention mechanisms can do this be having such direct channels.Such a model can use an encoder-decoder. For example, a translator which translates french to english; if the sentence in french has 10 words, that does not mean that the sentence in english will also have 10 words.True. During inference, batch normalization shifts and rescales independently each component of the input x according to statistics estimated during training.

14. What tends to happen if weights are ill conditioned at initialization?

Such a model can use an encoder-decoder. For example, a translator which translates french to english; if the sentence in french has 10 words, that does not mean that the sentence in english will also have 10 words.The gradients go towards zero in deeper layers.True. During inference, batch normalization shifts and rescales independently each component of the input x according to statistics estimated during training.Un-supervised. This is a big benefit as it allows to train on potentially huge corpora of data.

15. What is an activation map?

A channel produced by a convolution layer.True. During inference, batch normalization shifts and rescales independently each component of the input x according to statistics estimated during training.Un-supervised. This is a big benefit as it allows to train on potentially huge corpora of data.Such a model can use an encoder-decoder. For example, a translator which translates french to english; if the sentence in french has 10 words, that does not mean that the sentence in english will also have 10 words.

16. Is the backward pass more computationally expensive than the forward pass?

Yes, by a factor of about 2.Such a model can use an encoder-decoder. For example, a translator which translates french to english; if the sentence in french has 10 words, that does not mean that the sentence in english will also have 10 words.Un-supervised. This is a big benefit as it allows to train on potentially huge corpora of data.True. During inference, batch normalization shifts and rescales independently each component of the input x according to statistics estimated during training.

17. Name one characteristic of the t-Distributed Stochastic Neighbor Embedding (t-SNE).

It optimizes the SGD with the yis to maintain the distances to close neighbors of each point. This makes it easier to distinguish between clusters. Often used to reduce high dimensional data.Un-supervised. This is a big benefit as it allows to train on potentially huge corpora of data.Such a model can use an encoder-decoder. For example, a translator which translates french to english; if the sentence in french has 10 words, that does not mean that the sentence in english will also have 10 words.True. During inference, batch normalization shifts and rescales independently each component of the input x according to statistics estimated during training.

18. Is the autograd graph the same as the structure of the network?

Un-supervised. This is a big benefit as it allows to train on potentially huge corpora of data.No, they are related but may differ.Such a model can use an encoder-decoder. For example, a translator which translates french to english; if the sentence in french has 10 words, that does not mean that the sentence in english will also have 10 words.True. During inference, batch normalization shifts and rescales independently each component of the input x according to statistics estimated during training.

19. Variational Autoencoders

Such a model can use an encoder-decoder. For example, a translator which translates french to english; if the sentence in french has 10 words, that does not mean that the sentence in english will also have 10 words.True. During inference, batch normalization shifts and rescales independently each component of the input x according to statistics estimated during training.Provides a probabilistic manner for describing an observation in latent space.Un-supervised. This is a big benefit as it allows to train on potentially huge corpora of data.

20. What are two standard performance measures for image classification?

True. During inference, batch normalization shifts and rescales independently each component of the input x according to statistics estimated during training.Un-supervised. This is a big benefit as it allows to train on potentially huge corpora of data.1) The error rate, or conversely accuracy. 2) The balanced error rate (BER)Such a model can use an encoder-decoder. For example, a translator which translates french to english; if the sentence in french has 10 words, that does not mean that the sentence in english will also have 10 words.

21. Name some techniques which help the training of very deep architectures.

True. During inference, batch normalization shifts and rescales independently each component of the input x according to statistics estimated during training.Such a model can use an encoder-decoder. For example, a translator which translates french to english; if the sentence in french has 10 words, that does not mean that the sentence in english will also have 10 words.Un-supervised. This is a big benefit as it allows to train on potentially huge corpora of data.1) Rectifiers: prevent the vanishing of the gradient 2) Dropout: force a distributed representation 3) Batch Normalization: dynamically maintain the statistics of the activations 4) smart initial conditions

22. What do bias and variance quantify in a model?

Such a model can use an encoder-decoder. For example, a translator which translates french to english; if the sentence in french has 10 words, that does not mean that the sentence in english will also have 10 words.The bias quantifies how much the model fits on average and the variance quantifies how much the model changes across different datasets.Un-supervised. This is a big benefit as it allows to train on potentially huge corpora of data.True. During inference, batch normalization shifts and rescales independently each component of the input x according to statistics estimated during training.

23. What are the two main components of a GAN network?

Un-supervised. This is a big benefit as it allows to train on potentially huge corpora of data.A generator, which generates data and a discriminator, which predicts whether the data is from the generator or the true data set.Such a model can use an encoder-decoder. For example, a translator which translates french to english; if the sentence in french has 10 words, that does not mean that the sentence in english will also have 10 words.True. During inference, batch normalization shifts and rescales independently each component of the input x according to statistics estimated during training.

24. When are penalties most useful?

25. In autoregressive models, why are the best results achieved with cross-entropy?

Such a model can use an encoder-decoder. For example, a translator which translates french to english; if the sentence in french has 10 words, that does not mean that the sentence in english will also have 10 words.True. During inference, batch normalization shifts and rescales independently each component of the input x according to statistics estimated during training.This is due in large part to the ability of categorical distributions and cross-entropy to deal with exotic posteriors, in particular multi-modal. (A multimodal posterior is when conflicting interpretations of x exist)Un-supervised. This is a big benefit as it allows to train on potentially huge corpora of data.

⚡ Recently practiced quizzes in this topic

Live quiz activity

Machine Learning Test Machine Learning: Recommendation Systems Questions Machine Learning 101 Practice Test: Linear Regression Machine Learning Basics Knowledge Test Machine Learning 101 Practice Test: Fundamental Theorem of PAC Learning Machine Learning 101 Practice Test: Kernels And Kernel Trick Machine Learning 101 Practice Test: K-Nearest Neighbor Algorithm and Nearest Neighbor Analysis Machine Learning 101 Practice Test: Neural Networks in Machine Learning Machine Learning 101 Practice Test: Decision Trees Machine Learning 101 Practice Test: Version Spaces, Find-S Algorithm And Candidate Elimination Algorithm

25 Questions

❤ If you liked Fatskills, consider supporting us by checking out The Life Manuals You Never Got.

About | Explore | User Guide | Topics | Subjects | Doubt Solver | Career Aptitude Test | Answers | Free Tools | OSHA Basics Quiz | What Should We Know?
Privacy | Terms |

Without work one finishes nothing. - Ralph Waldo Emerson
© 2026 Fatskills.com

All trademarks, logos and brand names are the property of their respective owners. All company, product and service names used in this website are for identification purposes only. Use of these names, trademarks and brands does not imply endorsement.

25 Questions

❤ If you liked Fatskills, consider supporting us by checking out The Life Manuals You Never Got.

About | Explore | User Guide | Topics | Subjects | Doubt Solver | Career Aptitude Test | Answers | Free Tools | OSHA Basics Quiz | What Should We Know? Privacy | Terms |

Without work one finishes nothing. - Ralph Waldo Emerson© 2026 Fatskills.com

All trademarks, logos and brand names are the property of their respective owners. All company, product and service names used in this website are for identification purposes only. Use of these names, trademarks and brands does not imply endorsement.

About | Explore | User Guide | Topics | Subjects | Doubt Solver | Career Aptitude Test | Answers | Free Tools | OSHA Basics Quiz | What Should We Know?
Privacy | Terms |

Without work one finishes nothing. - Ralph Waldo Emerson
© 2026 Fatskills.com