Amazon Yapay Zeka Testi

Testi çözdüğünüzde cevaplar ve açıklamalar görünür.

1. How does Amazon Bedrock differ from Amazon SageMaker AI at a high level?
   • Bedrock offers managed access to foundation models, while SageMaker AI supports building, training and deploying models
   • They are identical products with different names
   • Bedrock is a physical GPU and SageMaker is a database
   • SageMaker AI is Amazon’s home voice assistant
2. What should a developer verify before relying on a Bedrock model feature?
   • That every foundation model is available in every region
   • That the model and feature are available for the account, region and chosen configuration
   • That the feature works without permissions
   • That model access guarantees a particular output quality
3. Which statement about data handling across AWS AI services is supported by the guide?
   • A managed service removes every data responsibility
   • All AWS services use one identical data policy
   • Storage, logging, retrieval and model calls can have separate controls to review
   • Permission settings never affect model requests
4. What does using a managed AI service change for the application owner?
   • It guarantees legal compliance for every workflow
   • It removes the need to validate model outputs
   • It means no application configuration is needed
   • It reduces some infrastructure work but does not remove responsibility for authorized access and data use
5. Which deployment record helps reproduce an AWS AI evaluation?
   • The endpoint, model version and configuration used
   • Only the service’s marketing name
   • The user’s browser theme
   • A copy of unrelated training logs
6. What should a team include when comparing the cost of an AWS AI workflow?
   • Only the model’s parameter count
   • Storage, data transfer, failures, throughput and operational behavior
   • Only the first successful request
   • The number of tabs open in the console
7. Why can a model offered through Bedrock differ operationally from using its provider’s direct API?
   • The underlying model name makes all terms identical
   • Bedrock guarantees the provider’s exact interface
   • The managed platform can have different integration details, access conditions and configuration
   • Using a managed endpoint removes model versioning
8. Before production rollout, which test best follows the guide’s service-selection advice?
   • Assume the public feature list applies to every account
   • Test only a model’s training accuracy
   • Skip tests if the service is managed
   • Evaluate representative tasks and failure handling in the intended account and region