Quiz Space

Deep Learning for Computer Vision · Quiz 2 · 4 Aug 2024 · May 2024 term

Question 26: KL divergence is not a distance metric because \text{KL}…

Question 26

+0.5 marksNumerical answer

KL divergence is not a distance metric because KL(P∥Q)\text{KL}(P\|Q) is (a) __________ to KL(Q∥P)\text{KL}(Q\|P). And KL divergence is always ⩾\geqslant (b) __________.
Suppose P∼Bernoulli(0.8)P \sim \text{Bernoulli}(0.8) and Q∼Bernoulli(0.2)Q \sim \text{Bernoulli}(0.2) then
(c) KL(P∥Q)=\text{KL}(P\|Q) = __________ nats and (d) KL(Q∥P)=\text{KL}(Q\|P) = __________ nats.

Hint: You can use these values of natural log: ln⁡(2)=0.69\ln(2) = 0.69, ln⁡(4)=1.38\ln(4) = 1.38, ln⁡(0.6)=−0.510\ln(0.6) = -0.510, ln⁡(8)=2.079\ln(8) = 2.079, ln⁡(0.4)=−0.91\ln(0.4) = -0.91,
Answer upto two decimal places.

Based on the above data, answer the given subquestions.

Enter the correct answer for (d)

Show answer

Correct answer: 0.83

See the whole paperTake it as a mock test

Question 26 of 29 in the IIT Madras BS Deep Learning for Computer Vision (Deep Learning for Computer Vision) Quiz 2 paper sat on 4 Aug 2024, in the May 2024 term (IIT M DEGREE AN EXAM QDB2 4 Aug 2024). It carries 0.5 marks.

More questions from this paper

  1. Q1Which one of the below statements is incorrect?
  2. Q2Which one of the following statements is false?
  3. Q3Match the following:
  4. Q4Identify the true statement regarding different loss functions:
  5. Q5In the context of Variational Autoencoders, the idea of variational inference solves the problem of
  6. Q6Which of the following is incorrect for the reparametrization trick used in a Variational AutoEncoder (VAE)?
  7. Q7In an image classification problem, which of the following is the quantity of interest in a generative model?
  8. Q8Figure question
  9. Q9Figure question
  10. Q10Which one of the following statements is true?
  11. Q11Which one of the following statements is false?
  12. Q12Consider two 10×10 bounding boxes(one on the upper left and one of the lower right) in an images with an overlapping re…
  13. Q13A four-dimensional input vector x = [5, 3,−1, 2] is passed to a hidden layer with a single neuron and an activation fun…
  14. Q14Figure question
  15. Q15Figure question
  16. Q16Figure question
  17. Q17Consider a Faster R-CNN object detection framework, which is being trained with a mini-batch size of 2. The RPN produce…
  18. Q18The feature embeddings obtained from a neural network for an anchor sample, positive sample and negative sample are as …
  19. Q19A Siamese network is trained for face verification. The face image to be verified is passed through the first branch an…
  20. Q20z1=_
  21. Q21z2=
  22. Q22z3=_
  23. Q23KL divergence is not a distance metric because \text{KL}(P|Q) is (a) to \text{KL}(Q|P). And KL divergence is always \ge…
  24. Q24KL divergence is not a distance metric because \text{KL}(P|Q) is (a) to \text{KL}(Q|P). And KL divergence is always \ge…
  25. Q25KL divergence is not a distance metric because \text{KL}(P|Q) is (a) to \text{KL}(Q|P). And KL divergence is always \ge…
  26. Q27Consider generative and discriminative models for a classification task. Generative models learn the data-generating me…
  27. Q28Which of the following are true:
  28. Q29Which of the following is true regarding Hard Attention and Soft Attention?