Free NVIDIA NCA-GENL Actual Exam Questions - Question 11 Discussion
performance of the model using A/B testing. What is the rationale for using A/B testing with deep
learning model performance?
A imo, since A/B testing is about comparing real user impact, not model robustness or latency.
A, because it’s about comparing user outcomes between two model versions.
It’s A because A/B testing focuses on comparing user responses to different model versions, so you can see which one actually improves recommendations in practice, not just in theory.
A/D? A seems solid since A/B testing is about direct comparison, but D also makes sense because latency can impact user experience, and measuring it could be part of evaluating model performance. Even if not the main goal, collecting latency data during A/B testing is practical. Options B and C feel off since B talks about rationale from designers—which isn’t really what A/B testing is for—and C talks about robustness, which is more about training and validation than A/B testing itself.
Makes sense to go with A since it directly compares two models’ user impact. A
I think B and C can be ruled out since A/B testing is more about live user impact rather than internal model details or input variation. D mentions latency, but that’s just one aspect, not the core reason for A/B testing here.
It’s A because A/B testing’s main goal is to compare user responses to different models in a live environment, which really shows which one works best in practice, not just in theory. D focuses too much on latency, which is just one part of the picture.
D imo, latency comparison is a key part of evaluating model performance in real scenarios. While A is true, measuring response time directly impacts user experience, which A/B testing captures well.
A. Besides just comparing which model performs better, A/B testing also simulates real user interaction, which is crucial for recommendation systems. Options B, C, and D feel off because they focus on technical explanations or latency, which are not the main point of A/B testing here. It's really about measuring actual user response to different versions.
I think the answer is A. A/B testing is mainly about comparing two versions directly to see which one performs better in a real setting. The rest don't seem to fit the core purpose of A/B testing. Anyone else agree?