Skip to content

Topic Drill 1.1.7: Machine Learning Development Frameworks ​

About this drill

LO AI-1.1.7 (K2): Summarize the functionality provided by ML development frameworks.
Lesson: Ch 1, section 1.1.7

This drill covers every point in section 1.1.7:

  • The five functions: data handling, model building, training and optimization, evaluation, deployment (in some frameworks)
  • Abstraction levels: lower-level vs. higher-level APIs
  • General-purpose vs. specialized frameworks
  • The seven selection factors

Topic Drill: 1.1.7

Q1. Before training, a data scientist uses the framework to load raw records, remove malformed entries and convert dates into a consistent format. Which framework functionality is this?

AModel building
BData handling
CEvaluation
DDeployment

Q2. A developer specifies that the model is a neural network with four hidden layers and decides how the layers are connected. Which framework functionality supports this?

AData handling
BEvaluation
CDeployment
DModel building

Q3. The framework repeatedly adjusts the model's internal parameters based on the training data, and can spread the work across several machines or fine-tune a pretrained model. Which functionality is this?

ATraining and optimization
BEvaluation
CData handling
DModel building

Q4. A team measures the error rate of its ML regression model on data the model has not seen before. Which framework functionality is being used?

ADeployment
BTraining and optimization
CEvaluation
DData handling

Q5. The trained model must be converted into a format that runs inside a mobile app. Which statement about this framework functionality is CORRECT?

AEvery ML development framework provides it
BSome frameworks provide deployment capabilities, such as converting the model for web, mobile, edge or embedded targets
CDeployment is part of data handling
DDeployment is only possible for models built with a higher-level API

Q6. A team with limited coding experience must build a working prototype within two weeks. Which framework characteristic suits them BEST?

AA lower-level API that gives maximum control over model building
BA specialized framework for speech recognition
CA framework designed for resource-constrained embedded devices
DA higher-level API that simplifies model creation, at the cost of fewer customization options

Q7. Which TWO of the following are selection factors for ML development frameworks listed in the syllabus?

Select 2 options.

AEcosystem maturity
BThe number of stars the framework has on code-hosting sites
CThe framework vendor's country of origin
DLevel of (community) support
EPopularity among recruiters

Q8. A framework focuses exclusively on speech recognition. How does the syllabus classify such a framework?

AAs a specialized framework, as opposed to a general-purpose one
BAs a lower-level API framework
CAs AI-specific hardware
DAs an AI as a Service offering

0 / 8 answered

← 1.1.6 Development and Hosting · Next drill: 1.1.8 Regulations and Standards →

Unofficial study guide for the ISTQB® Certified Tester AI Testing (CT-AI) exam. Not affiliated with International Software Testing Qualifications Board (ISTQB®).