Scribd
Upload
86 views1 page

Week 01 Assignment 01

asign

Uploaded by

Uma Maheshwar
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF or read online on Scribd
86 views1 page

Week 01 Assignment 01

asign

Uploaded by

Uma Maheshwar
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF or read online on Scribd
You are on page 1/ 1
‘Avout the Course Unit 11 - Week 9: ‘Course outtine ow to acess te porta Wook 4: fe taca 4: GongeNt Optmaer: Marti, nimser aca 0: opine: ooo Greet NAC) numer tec 4: opine: esa Optsee case 4: Opa sirp Asani rs Asem Oaimset Assignment 9 ‘The de date forsubmiting thi asgnment hs posse Due on 2019-10-02, 23:59 IST, Gradient ofa particular dimension points to the same direction, What can you say about the ‘momentum parameter, y. Choose the correct option. 14. The momentum term increasee b. The momentum term decreases Cannot comment 4 Itremains the same on reves caret decent newer: 2) Comment on the learning rat of Adagrad. Choose the correct option, Learning rates adaptive Learning rate increases for each time step Learning rate remains the same for each update None of the above a some iaincane. ‘Aecentes Aree: 5) Adagrad has ts own limitations. Can you choose that limitation from the following options? a. Accumulation ofthe positive squared gradients in the denominator b. Overshooting minima © Learning rate increases thus hindering convergence and cause the loss function to fluctuate around the minimum or even to diverge 4. Getting trapped in local minima ‘Accept Answers: ©) what isthe full form of RMSProp? Retain Momentum Propagation Round Mean Square Propagation Root Mean Square Propagation ‘None of the above ‘Accent Anewers: 5) RMSProp resolves the imitation af which optimizer? Adagrad Momentum Both a and Neither a norb ‘Accented Aree: © Which ofthe following statement is true? ‘2 Gradient update rule for Momentum optimizer and Nesterov Accelerated Gradient (NAG) optimizer are same b, Momentum optimizer and Nesterov Accelerated Gradient (NAG) optimizer perform differently irrespective of learning rates c. Possibility of oscillations is lower in Nesterov Accelerated Gradient (NAG) than Momentum optimiser 1d. None of the above ede anoer wince Seon serpin: 7 The following i the equation of update vector for monet optimizer: veeyves +11 Vos(8) What isthe range of 7? a. Oand 1 b. >0 te ne anoweraincaeet, Sear ‘Accented Aree: ©) Why itis at all required to choose different learning rates for different weights? To avoid the problem of diminishing learning rate ‘To avoid overshooting the optimum point To reduce vertical oscillations while navigating the optimum point This would aid to reach the optimum point faster ne nena incomet ‘econ Anewors ® What isthe major drawback of setting large leering rate for updating weight parameter for Gradient Descent? Slower convergence Struck in local minima Overshoots optimum point None of the above ‘ecente Answers: 79) Fora smaller magnitude gradient descent, what should be the suggested learning rate for updating the weights? small Large cannot comment same learning rate for small and large gradient magnitudes seven tars: 2pointe pointe 2pointe 2 points 2aints pons 2pointe pons pointe 2 points

You might also like