CS 221 exam-2017 Stanford University

Document Content and Description Below

CS221 Exam Solutions CS221 November 28, 2017 Name: | {z } by writing my name I agree to abide by the honor code SUNet ID: Read all of the following information before starting the exam: • Thi... s test has 3 problems and is worth 150 points total. It is your responsibility to make sure that you have all of the pages. • Keep your answers precise and concise. Show all work, clearly and in order, or else points will be deducted, even if your final answer is correct. • Don’t spend too much time on one problem. Read through all the problems carefully and do the easy ones first. Try to understand the problems intuitively; it really helps to draw a picture. • You cannot use any external aids except one double-sided 81 2" x 11" page of notes. • Good luck! Problem Part Max Score Score 1 a 15 b 15 c 20 2 a 10 b 10 c 10 d 10 e 10 3 a 15 b 10 c 15 d 10 Total Score: + + = 11. Learning (50 points) a. (15 points) [Generalization] For problems (i){(iii), circle one of the bolded options. (i) [3 points] To decrease training error, would you want more or less data? Solution Less. Fewer data points are easier to fit. (ii) [3 points] To decrease training error, would you want to add or remove features? Solution Add. More features makes it easier to fit the data. (iii) [3 points] To decrease training error, would you want to make the set of hypotheses smaller or larger? Solution Larger. More hypotheses makes it easier to fit the data. 2(iv) [3 points] If a learning algorithm generalizes very well, what does this say about the training error and the test error? Your answer should be one sentence. Solution The training and the test errors are approximately equal. Note that strictly speaking, we do not require the test error to be small. (v) [3 points] In class, we talked about dividing the data into train, validation, and test sets. Give a way that you might use the validation set. Your answer should be one sentence. Solution The validation set can be used to (1) tune hyperparameters, (2) choose features, (3) choose hypothesis class, or other decisions where you want to estimate the test error, without using the test error [Show More]

Last updated: 2 years ago

Preview 1 out of 25 pages

Buy Now

Instant download

We Accept: