Machine Learning Questions - Study24x7
Social learning Network

Warning: include(./views/auth.php): failed to open stream: Permission denied in /var/www/html/live/loginRightSlider.php on line 18

Warning: include(): Failed opening './views/auth.php' for inclusion (include_path='.:/usr/share/php') in /var/www/html/live/loginRightSlider.php on line 18
machinelearningquestions Cover image
machinelearningquestions
Machine Learning Questions

  • Warning: Use of undefined constant Followers - assumed 'Followers' (this will throw an Error in a future version of PHP) in /var/www/html/live/makepage.php on line 1760
    Followers
  • Latest Feeds
  • Articles
  • Question
  • Course
  • About
  • Review
Interests
Web Development

Warning: count(): Parameter must be an array or an object that implements Countable in /var/www/html/live/makepost.php on line 52
7 followers study24x7 01 Jan 1970 05:30 AM study24x7 study24x7

What is the primary role of the discriminator in a GAN?

A

Generate images

B

Judge real vs fake images

C

Encode images

D

Normalize noise

study24x7
Write a comment

Warning: count(): Parameter must be an array or an object that implements Countable in /var/www/html/live/makepost.php on line 52
7 followers study24x7 01 Jan 1970 05:30 AM study24x7 study24x7

What is an epoch?

A

A single sample

B

One complete pass through training data

C

One batch

D

One optimizer step

study24x7
Write a comment

Warning: count(): Parameter must be an array or an object that implements Countable in /var/www/html/live/makepost.php on line 52
7 followers study24x7 01 Jan 1970 05:30 AM study24x7 study24x7

Why are GELU activations popular in Transformers?

A

Less compute

B

Smooth non-linearity improves gradient flow

C

Integer weights

D

Fixed embedding size

study24x7
Write a comment

Warning: count(): Parameter must be an array or an object that implements Countable in /var/www/html/live/makepost.php on line 52
7 followers study24x7 01 Jan 1970 05:30 AM study24x7 study24x7

What is “weight decay” used for?

A

Speeding up training

B

Penalizing large weights to reduce overfitting

C

Reducing model size

D

Increasing learning rate

study24x7
Write a comment

Warning: count(): Parameter must be an array or an object that implements Countable in /var/www/html/live/makepost.php on line 52
7 followers study24x7 01 Jan 1970 05:30 AM study24x7 study24x7

What do activation functions introduce in a neural network?

A

Linearity

B

Non-linearity

C

More parameters

D

More memory

study24x7
Write a comment

Warning: count(): Parameter must be an array or an object that implements Countable in /var/www/html/live/makepost.php on line 52
7 followers study24x7 01 Jan 1970 05:30 AM study24x7 study24x7

What is a tokenizer’s job in an LLM pipeline?

A

Train embeddings

B

Break text into model-readable units

C

Generate weights

D

Reduce GPU memory

study24x7
Write a comment

Warning: count(): Parameter must be an array or an object that implements Countable in /var/www/html/live/makepost.php on line 52
7 followers study24x7 01 Jan 1970 05:30 AM study24x7 study24x7

What is the key benefit of rotary positional embeddings (RoPE)?

A

Reduces vocabulary

B

Helps with long-context extrapolation

C

Removes attention computation

D

Improves tokenization speed

study24x7
Write a comment

Warning: count(): Parameter must be an array or an object that implements Countable in /var/www/html/live/makepost.php on line 52
7 followers study24x7 01 Jan 1970 05:30 AM study24x7 study24x7

What does PPO do in RLHF?

A

Compresses models

B

Stabilizes reinforcement learning updates

C

Tokenizes text

D

Reduces model layers

study24x7
Write a comment

Warning: count(): Parameter must be an array or an object that implements Countable in /var/www/html/live/makepost.php on line 52
7 followers study24x7 01 Jan 1970 05:30 AM study24x7 study24x7

Why is multi-head attention used instead of single-head?

A

Faster training

B

To capture different types of relationships

C

To reduce embeddings

D

To increase token limit

study24x7
Write a comment

Warning: count(): Parameter must be an array or an object that implements Countable in /var/www/html/live/makepost.php on line 52
7 followers study24x7 01 Jan 1970 05:30 AM study24x7 study24x7

What is the main purpose of the Query vector in self-attention?

A

Generate output tokens

B

Match relevance with Key vectors

C

Normalize embeddings

D

Reduce loss

study24x7
Write a comment

Warning: count(): Parameter must be an array or an object that implements Countable in /var/www/html/live/makepost.php on line 52
7 followers study24x7 01 Jan 1970 05:30 AM study24x7 study24x7

What is model quantization used for?

A

Increase accuracy

B

Reduce model size by lowering precision

C

Increase training time

D

Add more parameters

study24x7
Write a comment

Warning: count(): Parameter must be an array or an object that implements Countable in /var/www/html/live/makepost.php on line 52
7 followers study24x7 01 Jan 1970 05:30 AM study24x7 study24x7

What is attention head in multi-head attention?

A

Tokenizer

B

Independent attention mechanism

C

GPU thread

D

Decoder layer

study24x7
Write a comment

Warning: count(): Parameter must be an array or an object that implements Countable in /var/www/html/live/makepost.php on line 52
7 followers study24x7 01 Jan 1970 05:30 AM study24x7 study24x7

What is a token in LLMs?

A

Single character only

B

Building block used for model input

C

GPU core

D

Image pixel

study24x7
Write a comment

Warning: count(): Parameter must be an array or an object that implements Countable in /var/www/html/live/makepost.php on line 52
7 followers study24x7 01 Jan 1970 05:30 AM study24x7 study24x7

What is the main advantage of diffusion models?

A

Low parameter count

B

Stable training and higher image quality

C

No GPU needed

D

Faster than GANs always

study24x7
Write a comment

Warning: count(): Parameter must be an array or an object that implements Countable in /var/www/html/live/makepost.php on line 52
7 followers study24x7 01 Jan 1970 05:30 AM study24x7 study24x7

What does KL divergence measure in VAEs?

A

Distance between two probability distributions

B

Latency

C

Model size

D

Accuracy

study24x7
Write a comment

Warning: count(): Parameter must be an array or an object that implements Countable in /var/www/html/live/makepost.php on line 52
7 followers study24x7 01 Jan 1970 05:30 AM study24x7 study24x7

What does an autoencoder mainly learn?

A

Labels

B

Compressed representation of data

C

Random noise

D

Token probabilities

study24x7
Write a comment

Warning: count(): Parameter must be an array or an object that implements Countable in /var/www/html/live/makepost.php on line 52
7 followers study24x7 01 Jan 1970 05:30 AM study24x7 study24x7

What is the benefit of mini-batch gradient descent?

A

Slowest training

B

Most memory required

C

Balance between speed and stability

D

Never converges

study24x7
Write a comment

Warning: count(): Parameter must be an array or an object that implements Countable in /var/www/html/live/makepost.php on line 52
7 followers study24x7 01 Jan 1970 05:30 AM study24x7 study24x7

What is “negative sampling” in Word2Vec?

A

Removing duplicates

B

Efficiently training embeddings with fewer calculations

C

Increasing vocabulary

D

Tokenizing text

study24x7
Write a comment

Warning: count(): Parameter must be an array or an object that implements Countable in /var/www/html/live/makepost.php on line 52
7 followers study24x7 01 Jan 1970 05:30 AM study24x7 study24x7

Why are GPUs preferred for deep learning?

A

More RAM

B

Fast parallel matrix operations

C

Larger storage

D

Fewer cores

study24x7
Write a comment

Warning: count(): Parameter must be an array or an object that implements Countable in /var/www/html/live/makepost.php on line 52
7 followers study24x7 01 Jan 1970 05:30 AM study24x7 study24x7

What is the purpose of beam search?

A

Random token selection

B

Explore multiple sequences during decoding

C

Remove noise

D

Reduce training time

study24x7
Write a comment
Ratings
0.0
out of 5
0 Ratings
5 study24x7
 
0.0
4 study24x7
 
0.0
3 study24x7
 
0.0
2 study24x7
 
0.0
1 study24x7
 
0.0
Related Pages