Which ONE of the following options BEST describes clustering?
AClustering is classification of a continuous quantity.
BClustering is done without prior knowledge of output classes.
CClustering is supervised learning.
DClustering requires you to know the classes.
What requirements are given in text documents, which ONE of the following is the BEST way to generate test cases from these requirements?
ANatural language processing on textual requirements
BAnalyzing source code for generating test cases
CGUI analysis using computer vision
DMachine learning on logs of execution
Which ONE of the following statements about the hardware used to implement ML systems is MOST likely to be correct?
ASpecialist hardware is necessary for ML
BLess bits are required for hardware supporting ML
CSupport for complex operations is required
DA higher clock speed is required
In which ONE of the following situations would an ML model be MOST effective at determining the criticality of new defects?
AA new application which is in the early stages of the first test cycle
BAn old application with lots of defect records but a brand new development and test team
CAn old application where defect records are linked to failed tests and production incidents
DAn old application with few critical defect records and many non-critical defect records
Which ONE of the following would be the MOST effective input to an AI-based defect prediction tool?
ADevelopers associated with previous code changes
BCyclomatic complexity
CLines of code changed
DCommit size
Which ONE of the following options is a technology used to implement AI?
AGenetic algorithms
BAutonomy
CClassification
DReinforcement learning
Which ONE of the below is MOST likely to indicate a problem with underfitting in an ML model?
AThe model is vulnerable to adversarial attacks
BThe model fails to generalize on new data
CThe model uses a large amount of resources to make a prediction
DThe model is inaccurate on data similar to the training data
Which ONE of the below is NOT an input to an activation function in a neural network?
ABias
BActivation values from the previous layer
CActivation values from the current layer
DThe weights assigned to the connections
A test engineer is planning the best functional performance metrics to evaluate an unsupervised learning model. The model groups data points based on their similarity. The test engineer wants to measure how similar the data points in each group actually are.
Which is the MOST likely metric they should use?
AROC
BIntra-cluster
CAUC
DRecall
Which ONE of the following options for a test basis would give the LEAST coverage when using AI-based test generation?
AA test model describing the application functionality
BA list of possible web pages that comprise the application
CA pseudo-oracle
DAn XML schema
Which ONE of the following would be the LEAST effective input to an AI-based test optimization process?
APreviously failing tests
BDefect reports
CTest environment downtime
DSource control data
You are a test manager planning testing for an invoice financing company. The company buys unpaid invoices from companies and provides them with immediate cash.
The company is replacing their existing conventional system, which takes company accounting records as inputs, with an ML system that classifies each invoice for sales as something that should, or should not be bought. Significant historical production data is available. It is important that invoices are not bought incorrectly.
Which ONE of the following test techniques would be MOST appropriate for you to plan for system testing?
AA/B testing
BBack-to-back testing
CMetamorphic testing
DExperience-based testing
Which ONE of the following statements is MOST true about black-box adversarial testing?
AThe algorithm used for the model is known to the tester
BThe approach relies on the transferability of the attacks
CThe training data is manipulated
DCombinatorial testing techniques are used
You have been asked for your opinion on the ML approach to be used for a new age verification system. A training dataset is available with many pictures of faces and the subject's age.
When new users sign-up to a mobile app, they have to take a picture of their face. If the system believes they are under 18 years old it will ask them to provide a picture of their passport to prove their age.
Which ONE of the following approaches do you expect to be MOST likely to succeed?
AUnsupervised learning that identifies clusters in training data that relate to people’s ages.
BSupervised learning classification system that is re-trained based on new data, including whether they later provided ID that showed they were over 18.
CSupervised learning regression system that is re-trained based on new data, including whether they provided ID that showed their age.
DReinforcement learning classification system with a reward function of correct behavior.
Which ONE of the following statements BEST describes a testing challenge that specifically applied to a self-learning system?
AWhen systems change themselves the results of previously passing tests may change
BIt is necessary to test whether the system will relinquish control to a human at the right time
CExternal data sources might be required to ensure that the system is unbiased
DIn can be difficult to explain the link between test inputs and outputs
Which ONE of the following statements about a system MOST describes an autonomous system?
AA loan approval system that can continue to process loan applications indefinitely, within a defined amount of aggregate credit, until an operator reduces the amount of aggregate credit available.
BA self-driving car that automatically stops the car if the driver is not responding to voice prompts, in order to ensure they are awake.
CA chatbot that learns the most effective responses to humans (in order to ensure the humans continue to converse) based on prior experiences.
DA fraud detection system that alerts operators when it sees a specific fraud risk associated with a transaction, in order to obtain human input to make a final decision.
A facial recognition system is being deployed at airports in order to scan passengers’ faces and compare them to a database of vaccinations, in order to identify unvaccinated passengers in a pandemic.
There are a number of components involved including cameras, a model to segment the image, and a model to identify the face and match it against a known photograph. It is important that there are few false negatives, and that passengers cannot subvert the system.
Which ONE of the following types of testing is the MOST appropriate options for the tests you would choose in system testing?
ATesting for concept drift
BAdversarial testing
CTesting for explainability
DTesting for transparency
Which ONE of the following requirements for an AI-based system is MOST likely to cause a challenge for testing?
AThe system should be able to accept natural language input on a variety of topics and respond correctly
BThe system should respond more quickly than the system it is replacing
CThe system should be 100% accurate with no tolerance for failure
DThe system should not require human intervention until a specific failure scenario is encountered
A data scientist is performing unsupervised learning on a set of financial records relating to previous loan applications, and trying to predict defaults on future loans. They are reporting poor functional performance because of data issues.
Which ONE of the below is LEAST likely to be a contributory factor?
AMissing records for some accounts
BMissing data relating to whether loans were previously granted and repaid
CInconsistent pre-processing of some records
DIrrelevant data included in the account records
Which ONE of the following statements is true about dynamic testing for inappropriate bias?
AIt can be necessary to obtain additional attributes about the data being processed
BReviewing the source of the training data can reveal inappropriate bias
CInappropriate bias only needs to be tested when protected characteristics such as race and gender are present in the inputs
DTesting should never be conducted in production
Which ONE of the following statements provides the BEST description of the AI effect?
AAs AI systems improve, the likelihood of reaching the technological singularity increases.
BAI systems get better over time as they are exposed to more data.
CThe systems used to play chess against human experts in the 1990’s are no longer considered AI.
DAI systems increasingly behave like humans.
Which ONE of the below statements BEST describes why test environments for autonomous systems might need to be different to other test environments?
ATools may be required to simulate extreme scenarios
BNon-determinism may need to be introduced into the environment
CAI-specific hardware may be required
DTools may be required to provide explanations of system behavior
Which of the below is NOT a characteristic that makes it difficult to use AI-based systems?
AComplexity
BLack of standards
CLack of robustness
DLack of transparency
Which ONE of the following is a factor associated with the test data that can create challenges specific to testing AI-based systems?
AObtaining data from public benchmark datasets
BObtaining data that has not been pre-processed by the system under test.
CManual creation of test data
DObtaining appropriate data from users
Which ONE of the below statements BEST describes how combinatorial testing can be applied to AI-based systems?
AEach neuron can be treated as a parameter for pairwise tests
BTwo variants of the system can be used and compared
CInputs to the system and environment factors can be considered parameters for pairwise tests
DCombinatorial testing cannot yet be applied to AI