Page 223 - Artificial Intellegence_v2.0_Class_12
P. 223
11. What is a loss function? Name any two loss functions.
Ans. Loss functions calculate how far an estimated value deviates from its true value.
If the forecasts are too far off from the actual findings, the loss function ill return a very large number. o
loss functions are—Mean Absolute Error, Log loss.
12. Why is MSE value never negative?
Ans. MSE value is never negative because the errors are always squared.
13. The RMSE of a good model should be less than ________________.
Ans. The RMSE of a good model should be less than 180.
14. What are hyperparameters? Name any two.
Ans. Hyperparameters are parameters whose values govern the learning process. They also determine the values
of model parameters learned by a learning algorithm.
For example: the ratio of train-test split, the loss function that the model will use.
15. What is the nature of the AI project cycle?
Ans. The AI project cycle is an iterative process.
16. Why is Problem Scoping difficult? Which methodology can help us?
Ans. coping a problem is difficult because e need to have a deeper understanding of it so that the image gets
clearer as we attempt to solve it. As a result, we employ the 4Ws Problem Canvas to assist us in identifying the
important factors associated with the problem.
17. Name few authentic sources of data.
Ans. • Government websites
• Devices such as cameras and sensors
• Purchases, transactions, registrations, and
• Other public surveys and records
18. Name any two AI development platforms that you have used/studied about.
Ans. IBM Watson, Anaconda
19. Why should data validation be carried out?
Ans. Human biases in picking test data might have a negative impact on the testing phase; thus, data validation is
critical.
20. List three points to be considered by the testing team during the Evaluation stage.
Ans. • Due to the sheer volume of data, performance testing is critical.
• If the AI solution requires data from other systems, systems integration testing is critical.
• All relevant subsets of training data, i.e., the data you will use to train the AI system should be included in
test data.
21. What is meant by test dataset? What is its another name?
Ans. he sample of data used to offer an unbiased evaluation of a final model fit on the training dataset is referred
to as the test dataset. The test data set is sometimes known as a holdout data set.
22. Why is a well-told story so compelling?
Ans. A well-told story with an inspiring narrative will always engage the audience across boundaries and cultures
since they have an impact that statistics alone cannot provide. Data can be convincing, but tales are far more
so. They alter our interactions with data, shifting it from a dry collection of "facts" to something that can be
interesting, engaging, thought-provoking, and encouraging change.
V i va V oce Q ue sti ons 221

