nerdexam
Google

CLOUD-DIGITAL-LEADER · Question #128

certain devices for cracks, rust, etc. Some of these issues are difficult to identify for a human and your company has seen increasing customer complaints -the customer has paid for an inspection…

The correct answer is A. Use AutoML Vision Edge models. AutoML Vision Edge models are the correct solution here because the key constraint is inconsistent mobile network connectivity in the field. Edge models are deployed directly on the device (e.g., a mobile phone or tablet), enabling inference to run locally without requiring a…

AI/ML Solution Design and Deployment

Question

certain devices for cracks, rust, etc. Some of these issues are difficult to identify for a human and your company has seen increasing customer complaints -the customer has paid for an inspection and the field agent said there was no problem, but it later turned out there actually was. The team has come up with a proposal to engage AI to identify issues. On evaluating the existing system, it is seen that the mobile phone network connection is not good or consistent. What solution can work for them?

Options

  • AUse AutoML Vision Edge models.
  • BUse the Rust programming language instead of Python to identify issues like rust.
  • CUse Cloud TPUs which will be able to do the analysis faster on the cloud. Thus re-sponses also
  • DUse TensorFlow to create custom models and deploy it as TensorFlow Lite mod-els.

How the community answered

(45 responses)
  • A
    84% (38)
  • B
    9% (4)
  • C
    2% (1)
  • D
    4% (2)

Explanation

AutoML Vision Edge models are the correct solution here because the key constraint is inconsistent mobile network connectivity in the field. Edge models are deployed directly on the device (e.g., a mobile phone or tablet), enabling inference to run locally without requiring a network connection. This eliminates dependency on cloud connectivity for real-time analysis. Option B is a joke answer (Rust language vs. rust detection). Option C (Cloud TPUs) requires cloud connectivity, which is unavailable. Option D (TensorFlow Lite) is similar in concept but AutoML Vision Edge specifically provides the training pipeline and edge-optimized model export for this exact use case - and is the Google-recommended approach.

Topics

#Edge AI#AutoML Vision#On-device Machine Learning#Network Connectivity

Community Discussion

No community discussion yet for this question.

Full CLOUD-DIGITAL-LEADER Practice