Google Professional Machine Learning Engineer Exam - Topic 6 Question 66 Discussion

Actual exam question for Google's Professional Machine Learning Engineer exam

Question #: 66
Topic #: 6

[All Professional Machine Learning Engineer Questions]

You work for a social media company. You want to create a no-code image classification model for an iOS mobile application to identify fashion accessories You have a labeled dataset in Cloud Storage You need to configure a training workflow that minimizes cost and serves predictions with the lowest possible latency What should you do?

ATrain the model by using AutoML, and register the model in Vertex Al Model Registry Configure your mobile
application to send batch requests during prediction.

BTrain the model by using AutoML Edge and export it as a Core ML model Configure your mobile application
to use the mlmodel file directly.

CTrain the model by using AutoML Edge and export the model as a TFLite model Configure your mobile application to use the tflite file directly

DTrain the model by using AutoML, and expose the model as a Vertex Al endpoint Configure your mobile application to invoke the endpoint during prediction.

Show Suggested Answer

Suggested Answer: D

Applying quantization to your SavedModel by reducing the floating point precision can help reduce the serving latency by decreasing the amount of memory and computation required to make a prediction. TensorFlow provides tools such as the tf.quantization module that can be used to quantize models and reduce their precision, which can significantly reduce serving latency without a significant decrease in model performance.

by Lai at Oct 22, 2023, 12:31 PM

Limited Time Offer

25%

Off

Get Premium Professional Machine Learning Engineer Questions as Interactive Web-Based Practice Test or PDF

Contribute your Thoughts:

Submit Cancel

Gladis

4 months ago

Batch requests? That's not efficient for mobile!

upvoted 0 times

...

Amie

4 months ago

Wait, can AutoML Edge really handle that?

upvoted 0 times

...

Brittni

4 months ago

Not sure about that, TFLite might be faster.

upvoted 0 times

...

Gladys

4 months ago

I agree, Core ML is perfect for mobile apps!

upvoted 0 times

...

Ayesha

4 months ago

Option B seems like the best choice for iOS.

upvoted 0 times

...

Gianna

5 months ago

I feel like using Vertex AI for model registration is a good idea, but I’m not sure if batch requests are the best approach for a mobile app.

upvoted 0 times

...

Ona

5 months ago

I’m a bit confused about whether to use batch requests or direct model invocation. I feel like direct invocation could be faster, but I need to double-check.

upvoted 0 times

...

Gail

5 months ago

I remember practicing a similar question where we had to minimize latency. I think exporting as a TFLite model might help with that.

upvoted 0 times

...

Cecily

5 months ago

I think using AutoML Edge could be the right choice since it’s designed for mobile applications, but I'm not entirely sure about the export formats.

upvoted 0 times

...

Alfred

5 months ago

This is a good opportunity to showcase my understanding of Google Cloud's AI/ML services. I'm confident I can analyze the question and select the most appropriate solution.

upvoted 0 times

...

Yasuko

5 months ago

Okay, I see the main tradeoffs are between using AutoML, AutoML Edge, and Vertex AI. I'll need to think through the pros and cons of each in terms of cost, latency, and model deployment for the mobile app.

upvoted 0 times

...

Janella

5 months ago

Hmm, I'm a bit unsure about the differences between the deployment options presented. I'll need to carefully review the details of each approach to determine the most cost-effective and low-latency solution.

upvoted 0 times

...

Miesha

5 months ago

This looks like a pretty straightforward question about deploying a no-code image classification model for a mobile app. I think I've got a good handle on the key considerations here.

upvoted 0 times

...

Merilyn

5 months ago

This looks like a straightforward question about the actions Server Protect can take when it detects an infected file. I'll need to carefully read through the options and select the three correct actions.

upvoted 0 times

...

Veronika

5 months ago

This question seems straightforward, but I want to make sure I understand the key tasks that text processing doesn't support well for requirements management.

upvoted 0 times

...

Carry

5 months ago

I think the answer is A. SAS Information Delivery Portal sounds like the right tool to surface different types of business and analytic content on the web.

upvoted 0 times

...

Tatum

5 months ago

This looks like a straightforward question on programming operators used in business process rules. I'll go through the options carefully and select the three that are supported.

upvoted 0 times

...

Gwen

10 months ago

Yo, I heard AutoML is like the easy mode of machine learning. Might as well just go with that and let the experts handle the hard stuff, am I right?

upvoted 0 times

Latanya

8 months ago

D) AutoML definitely simplifies the process. I think option A sounds like the way to go.

upvoted 0 times

...

Nobuko

8 months ago

C) Train the model by using AutoML Edge and export the model as a TFLite model Configure your mobile application to use the tflite file directly

upvoted 0 times

...

Lisha

8 months ago

B) Yeah, AutoML does make things easier. I'd go with that option for sure.

upvoted 0 times

...

Fredric

9 months ago

A) Train the model by using AutoML, and register the model in Vertex AI Model Registry Configure your mobile application to send batch requests during prediction.

upvoted 0 times

...

Coral

10 months ago

I'm gonna go with D. Vertex AI endpoint seems like the easiest option, even if it might cost a bit more.

upvoted 0 times

Lou

9 months ago

Yeah, it might be a bit more expensive, but it's worth it for the ease of use.

upvoted 0 times

...

Tran

9 months ago

That sounds like a good choice. It's convenient to use the Vertex AI endpoint for predictions.

upvoted 0 times

...

Cristy

9 months ago

D) Train the model by using AutoML, and expose the model as a Vertex AI endpoint Configure your mobile application to invoke the endpoint during prediction.

upvoted 0 times

...

Dusti

10 months ago

Hmm, I'm not sure. Option A with the batch requests might be a bit slower, but at least I don't have to worry about the model deployment.

upvoted 0 times

Alisha

9 months ago

I agree, Option C with the TFLite model could also be a good choice for low latency.

upvoted 0 times

...

Alisha

9 months ago

I think Option B might be better for minimizing latency, using the mlmodel file directly.

upvoted 0 times

...

Jani

10 months ago

Option A sounds good, batch requests can help with prediction speed.

upvoted 0 times

...

Gregoria

11 months ago

Option C looks good to me. Exporting the TFLite model and using it directly in the mobile app should give us the lowest possible latency.

upvoted 0 times

...

Cherilyn

11 months ago

I think option B is the way to go. Training with AutoML Edge and using the Core ML model directly on the mobile app sounds like the best approach to minimize cost and latency.

upvoted 0 times