Topic 1 - Introduction to AI Services

Question 1

What are the five main stages of a typical Machine Learning Workflow?

Answer 1

1. Data Preparation
2. Model Training
3. Model Evaluation
4. Model Deployment
5. Model Monitoring.

Question 2

Briefly explain the difference between the “API” and “DIY” implementations of a machine learning workflow in terms of infrastructure requirements

Answer 2

In the API implementation, the user does not need to manage infrastructure as they are using the service provider’s resources. In the DIY implementation, the user is responsible for providing and managing all necessary infrastructure. data training, and model development.

Question 3

In which implementation option is model development typically done by the user?

Answer 3

Model development is typically done by the user in the “Own Model” and “DIY” implementation options.

Question 4

What is an Application Programming Interface (API)?

Answer 4

An Application Programming Interface (API) is a set of definitions and protocols for building and integrating application software, allowing applications to communicate with each other.

Question 5

Describe the fundamental request/response paradigm that governs API communication.

Answer 5

The request/response paradigm involves a client sending a structured request to a server, and the server processing that request and returning a response with the requested information or result.

Question 6

What is the purpose of a status code in an API response, and what does a value of 200 typically indicate?

Answer 6

A status code indicates if the API request was successful or not. A value of 200 typically indicates a successful request.

Question 7

What data format is commonly used for transmitting data in API requests and responses, as seen in the Vision API examples?

Answer 7

JSON (JavaScript Object Notation) is the data format commonly used for transmitting data in API requests and responses.

Question 8

Give two examples of image analysis capabilities offered by the Google Cloud Vision API.

Answer 8

Two examples of image analysis capabilities offered by the Google Cloud Vision API are Label Detection and Text Detection. (Other valid answers include Document Text Detection, Facial Detection, Landmark Detection, Logo Detection, Image Properties).

Question 9

Looking at the provided JSON response example, how would you access the ‘score’ of the first ‘labelAnnotation’?

Answer 9

You would access the score using result[“responses”][0][“labelAnnotations”][0][“score”].

Question 10

Which implementation option generally has the highest associated risks?

Answer 10

The implementation option that generally has the highest associated risks is the “DIY” option.