Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                
Scribd Logo
Upload

Welcome to Scribd!

  • 1K views

    Assignment Week 11-Deep-Learning PDF

    Uploaded by

    ashish kumar
    This document contains a 10 question multiple choice quiz about deep learning topics covered in week 11 lectures, including triplet loss for face recognition, constraints in FaceNet's embedding function, transpose convolutions, challenges in facial recognition systems, differences between semantic and instance segmentation, semantic segmentation targets and losses, advantages of fully convolutional networks, and max unpooling. The questions test understanding of key concepts and each includes the correct answer and a reference to review the relevant lecture material.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd

    Assignment Week 11-Deep-Learning PDF

    Uploaded by

    ashish kumar
    1K views7 pages
    This document contains a 10 question multiple choice quiz about deep learning topics covered in week 11 lectures, including triplet loss for face recognition, constraints in FaceNet's embedding function, transpose convolutions, challenges in facial recognition systems, differences between semantic and instance segmentation, semantic segmentation targets and losses, advantages of fully convolutional networks, and max unpooling. The questions test understanding of key concepts and each includes the correct answer and a reference to review the relevant lecture material.

    Original Title

    Assignment Week 11-Deep-Learning.pdf

    Copyright

    © © All Rights Reserved

    Available Formats

    PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    This document contains a 10 question multiple choice quiz about deep learning topics covered in week 11 lectures, including triplet loss for face recognition, constraints in FaceNet's embedding function, transpose convolutions, challenges in facial recognition systems, differences between semantic and instance segmentation, semantic segmentation targets and losses, advantages of fully convolutional networks, and max unpooling. The questions test understanding of key concepts and each includes the correct answer and a reference to review the relevant lecture material.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    1K views7 pages

    Assignment Week 11-Deep-Learning PDF

    Uploaded by

    ashish kumar
    This document contains a 10 question multiple choice quiz about deep learning topics covered in week 11 lectures, including triplet loss for face recognition, constraints in FaceNet's embedding function, transpose convolutions, challenges in facial recognition systems, differences between semantic and instance segmentation, semantic segmentation targets and losses, advantages of fully convolutional networks, and max unpooling. The questions test understanding of key concepts and each includes the correct answer and a reference to review the relevant lecture material.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    of 7

    NPTEL Online Certification Courses

    Indian Institute of Technology Kharagpur

    Deep Learning
    Assignment- Week 11
    TYPE OF QUESTION: MCQ/MSQ
    Number of questions: 10 Total mark: 10 X 1 = 10
    ______________________________________________________________________________

    QUESTION 1:
    With context to triplet loss for face recognition, which of the following is true for easy negative
    example?

    a. Example whose embedding is near to anchor, and also belongs to same person
    as the anchor.
    b. Example whose embedding is near to anchor, but belongs to different person
    other than anchor.
    c. Example whose embedding is are far from the anchor, and belongs to same
    person as the anchor.
    d. Examples whose embedding is fare from anchor, and belongs to different person
    other than anchor.
    Correct Answer: d

    Detailed Solution:

    Please refer to the lectures of week 11.


    ______________________________________________________________________________

    QUESTION 2:
    In FaceNet which of the following constraint is imposed on the embedding function 𝑓(𝑥)?

    a. ‖𝑓(𝑥)‖22 =1
    b. ‖𝑓(𝑥)‖12 =1
    c. ‖𝑓(𝑥)‖22 <1
    d. ‖𝑓(𝑥)‖22 >1

    Correct Answer: a

    Detailed Solution:

    Please refer to the lectures of week 11.


    ______________________________________________________________________________
    NPTEL Online Certification Courses
    Indian Institute of Technology Kharagpur

    QUESTION 3:
    Suppose you have a 1𝐷 signal 𝑥 = [1,2,3,4,5] and a filter 𝑓 = [1,2,3,4], and you perform stride
    2 transpose convolution on the signal 𝑥 by the filter 𝑓 to get the signal 𝑦. What will be the
    signal 𝑦 if we don’t perform cropping?

    a. 𝑦 = [1,2,5,8,9,14,13,20,19,26,3,4]
    b. 𝑦 = [1,2,3,4,5,4,3,2,1]
    c. 𝑦 = [1,2,5,8,9,14,13,20,17,26,15,20]
    d. 𝑦 = [0,0,5,8,9,14,13,20,19,26,0,0]

    Correct Answer: c

    Detailed Solution:
    1 2 3 4 5

    1 1

    2 2

    1*3+2*1=5 3 1

    4*1+2*2=8 4 2

    3*2+1*3=9 3 1

    4*2+2*3=14 4 2

    3*3+1*4=13 3 1

    4*3+2*4=20 4 2

    3*4+1*5=17 3 1

    4*4+2*5=26 4 2

    15 3

    20 4

    ______________________________________________________________________________
    NPTEL Online Certification Courses
    Indian Institute of Technology Kharagpur

    QUESTION 4:
    What are the different challenges one face while creating a facial recognition system?

    a. Different illumination condition


    b. Different pose and orientation of face images
    c. Limited dataset for training
    d. All of the above

    Correct Answer: d

    Detailed Solution:

    Please refer to the lecture of week 11.

    ______________________________________________________________________________

    QUESTION 5:
    What is the difference between semantic segmentation and instance segmentation?

    a. Semantic segmentation treats multiple objects of the same class as a single


    entity. On the other hand, instance segmentation treats multiple objects of the
    same class as distinct individual objects.
    b. Both of them are same.
    c. Instance segmentation treats multiple objects of the same class as a single
    entity. On the other hand, semantic segmentation treats multiple objects of the
    same class as distinct individual objects.
    d. None of the above

    Correct Answer: a

    Detailed Solution:

    Instance segmentation treats multiple objects of the same class as distinct individual objects
    or instances.

    ____________________________________________________________________________
    NPTEL Online Certification Courses
    Indian Institute of Technology Kharagpur

    QUESTION 6:
    Which of the following is true for semantic segmentation?

    a. Semantic Segmentation can be considered as pixel wise classification problem.


    b. Semantic Segmented output has same dimension as the input image dimension.
    c. It has application in Autonomous driving, Industrial inspection, and Medical
    imaging analysis.
    d. All of the above

    Correct Answer: d

    Detailed Solution:

    Please refer to the lectures of week 11.

    ____________________________________________________________________________

    QUESTION 7:
    How we create the target for Semantic Segmentation?

    a. By one-hot encoding the class labels, essentially creating an output channel for
    each of the possible classes
    b. By labelling pixel of each class with different color
    c. By creating list of all possible classes with having the index of the pixel belonging
    to that particular class.
    d. None of the above

    Correct Answer: a

    Detailed Solution:

    Please refer to the lecture of week 11.

    ______________________________________________________________________________
    NPTEL Online Certification Courses
    Indian Institute of Technology Kharagpur

    QUESTION 8:
    Which of the following can be used as cost function for Semantic Segmentation?

    a. Pixel wise cross entropy Loss


    b. Mean Square Loss
    c. KL Divergence Loss
    d. None of the above

    Correct Answer: a

    Detailed Solution:

    Please refer to the lectures of week 11.

    ______________________________________________________________________________

    QUESTION 9:
    Fully Connected Convolutional network or FCN became one of the major successful network
    architectures. Can you identify what are the advantages of FCN which makes it a successful
    architecture for semantic segmentation?

    a. Larger Receptive Field


    b. Mixing of global feature
    c. Lesser computation required.
    d. All of the above

    Correct Answer: d

    Detailed Solution:

    Please refer to the lecture of week 11. It has a larger receptive field by using strided
    Convolution layer, also it mixes global feature and number of computations are reduced as
    we are down sampling the image resolution.

    ____________________________________________________________________________
    NPTEL Online Certification Courses
    Indian Institute of Technology Kharagpur

    QUESTION 10:
    In a Deep CNN architecture, the feature map before applying a max pool layer with (2x2) kernel
    is given bellow.
    12 6 15 9

    9 2 7 8

    14 2 17 6

    3 5 9 2

    After few successive convolution layers, the feature map is again up-sampled using Max Un-
    pooling. If the following is feature map before Max Unpooling layer. Then what will be the
    output of the Max Unpooling layer?
    5 6

    8 13
    NPTEL Online Certification Courses
    Indian Institute of Technology Kharagpur

    a.

    0 0 0 0

    5 0 0 6

    8 0 0 0

    0 0 13 0

    b.

    5 5 6 6

    5 5 6 6

    8 8 13 13

    8 8 13 13

    c.

    5 0 6 0

    0 0 0 0

    8 0 13 0

    0 0 0 0

    d. None of the above

    Correct Answer: d

    Detailed Solution:

    Please refer to the lectures of week 11.

    ______________________________________________________________________

    ______________________________________________________________________________

    ************END*******

    You might also like