Deep Learning
The questions can be answered in 2-4 sentences.
1. Give a method to fight vanishing gradient in fully-connected neural net- works. Assume we are using a network with Sigmoid activations trained using SGD.
2. You are designing a deep learning system to diagnose chest cancer through X-ray images. What do you think might be the most appropriate evaluation metric and why: Accuracy, Precision, Recall, F1 score.
Hi I would love to answer this question for you. I hope you would get a clear idea related to the same. So not wasting much time let's get started.
Ans.1 The most effective and the lates methodology to solve the problem of vanishing gradient in the fully connected neutral network is Residual network. This is the neutral network where the skip connection and residual connection are part of the networks architecture.
Ans.2 The most appropriate evaluation metric is Precision. The reason behind this selection is that the report must be precised so that the result mau be accurate as required.
I hope you have got a clear view on the solution, there are some other options for the above answers but these are best according to me. These questions may have more than one answer.
Please give thumbs up to an answer and do comment for any queries
Thanks and Happy to Help :)
HAPPY LEARNING
Get Answers For Free
Most questions answered within 1 hours.