AI Trailhead Questions

What can distort our understanding of artificial intelligence?

Which broad category would an AI system fit into if it’s used to determine the optimal price of an airline ticket?

A. Numeric prediction

B. Classification

C. Robotic navigation

D. Language processing

What limits programmers from handcrafting algorithms to perform tasks we associate with human intelligence?

A. Not enough memory in modern computers

B. Laws that prevent the creation of AI

C. The sheer number of rules to account for, many of which are unknown

D. Too little coffee, too little time

True or false: A database of business names, zip codes, and market value would be an example of structured data?

A. True

B. False

For AI training to be considered deep learning, what does its neural network need more of?

A. Nodes

B. Weights

C. Layers

D. Inputs

True or false: The values of weights and biases in a trained neural network usually have an obvious connection to the inputs.

A. True

B. False

What is it called when AI interprets everyday language?

A. Slang Translation

B. Text-to-task

C. Intention Prediction

D. Natural language processing

If you ask a generative AI what its favorite color is, and it responds “blue,” this is an example of what?

A. Sentience

B. Opinion

C. Prediction

D. Randomness

New AI model architecture and availability of extensive training data are two factors in the rapid improvement of generative AI. What’s the third?

A. Increased parallel computing power

B. AI optimizing AI code

C. Larger data storage capacity of servers

D. Faster satellite data connections

True or false: Developers must create their own large language models in order to add natural language processing to their applications.

A. True

B. False

What is natural language?

A. The root of all languages.

B. The ways humans communicate.

C. How computers speak to each other.

D. The language of plants.

In what ways have neural networks impacted NLP?

Which NLP technique uses the part of speech to more accurately find the root of a word?

A. Segmentation

B. Tokenization

C. Stemming

D. Lemmatization

What is the term for finding the underlying structure of text in NLP?

A. Parts of speech

B. Parsing

C. Morphology

D. Sentiment

What’s the difference between data-driven decision-making and traditional decision-making?

A. Data-driven decision-making is based on intuition and personal experience, while traditional decision-making is based on data analysis

B. Data-driven decision-making is based on data analysis, while traditional decision-making is based on intuition and personal experience

C. Data-driven decision-making and traditional decision making are the same

D. Traditional decision-making is becoming more important in modern organizations, while data-driven decision-making was prominent in the past

True or false: AI is a technology that enables machines to learn and perform tasks that would normally require human intelligence.

A. True

B. False

What are the three main classifications of data?

A. Numerical, categorical, and time-series

B. Structured, unstructured, and semi-structured

C. Quantitative, qualitative, and mixed

D. Tabular, image, and text

True or false: Data cleaning is the process of removing or correcting errors and inconsistencies in the data to improve its quality and accuracy.

A. True

B. False

What’s the main difference between machine learning and traditional programming?

A. Machine learning uses algorithms to learn from data and make predictions, while traditional programming follows a predefined set of rules

B. Machine learning requires less computation power compared to traditional programming

C. Machine learning and traditional programming are essentially the same, just different terminologies used in the field

D. Machine learning does not require any data, while traditional programming relies heavily on data

True or false: Ensuring data privacy and confidentiality is only important at the data collection stage.

A. True

B. False

What is GDPR?

A. A data protection law in United States

B. A data protection law in European Union

C. A data protection law in Australia

D. A data protection law protecting health data in the US

What does descriptive analytics tell leaders and stakeholders about their business?

A. What's going on based on data in the system

B. What's going to happen based on the past

C. What the gut feeling of the CEO is

D. What decision the leader should make

How do data analytics software tools help solve complex business scenarios?

A. They gamify insights to make work fun

B. They help scale data collection and visualize trends across a lot of data points

C. They automate technical architecture healing

D. They use artificial intelligence (AI) to make all critical decisions for the next strategic marketing campaign

True or false: Predictive analytics tools may be used to analyze security logs to anticipate and block bad web requests.

A. True

B. False

What is an advantage of doing data analytics in the cloud?

A. It doesn't require knowledge of the network and operating system

B. You can stop paying for infrastructure resources when they aren't needed

C. You get to experience data-center operations

D. A and B

E. B and C

True or false: Serverless technology enables organizations to focus on the analytics instead of server configuration.

A. True

B. False

The strength of the correlation of an r value of -0.52 can best be described as:

A. Very strong negative correlation

B. Very strong positive correlation

C. Modest negative correlation

D. Modest positive correlation

Which of these best describes correlation?

A. Pearson's correlation is a technique that can show causation

B. Pearson's correlation is a technique that can show how outliers affect relationships between variables

C. Pearson's correlation is a technique that can show whether and how strongly pairs of quantitative variables are related

D. Pearson's correlation is a technique that can show nonlinear relationshops

What does the r-squared value indicate?

A. Strength of the linear relationship between two quantitative variables

B. How well the model fits your observations

C. Whether the field adds or detracts from the significance of the entire model

D. How a value is related to the mean of a group of values

What's one definition of bias?

A. Decision made free of self-interest, prejudice, or favoritsm

B. Judgement based on preconceived notions or prejudices rather than the impartial evaluation of facts

C. The stage of being diverse and having variety

D. Impartial treatment without discrimination

Why are diverse teams important?

A. They enable you to create more inclusive products that meet the need of all your users, not just some of them

B. Only a few individuals are responsible for promoting ethics

C. They help you identify preconceived notions or assumptions that can be harmful for some users

D. A and C

E. B and C

Which of the following is a result of association bias?

A. Men being labeled as doctors and women being labeled as nurses in a dataset

B. An image of an orange cat being predicted to be a coyote because all the coyotes in the dataset were orange and none of the cats were orange

C. A person being denied a loan because the system inaccurately predicted they would be unable to repay it

D. A company hiring only candidates from a particular university because it currently has successful employees from the university

How can bias enter a system?

What are the advantages to holding a premortem?

A. Learn from past mistakes, celebrate successes, provide closure to a project

B. Identify potential risks in a project

C. Foster open communication with the team

D. A and C

E. B and C

Why is it important to regularly evaluate your data?

A. Societal values change over time

B. Your data model can "learn" unsavory information that skews the dataset

C. You can "set and forget" your data after evaluating it once

D. A and B

E. B and C

What are the risks of using generative AI?

A. Robots might take over Salesforce

B. You cannot limit the data the model gathers to generate content

C. You cannot create smaller models that are as accurate as the massive models available today

D. Toxicity and bias can cause harm at scale

Which of the following is an example of Salesforce's trusted approach to AI?

A. Hire robots to build privacy protections into products

B. Rely on customers to red team models

C. Scrape data off the web to train models

D. Red-team models before release to identify and address vulnerabilities

What is one advantage of using a diffusion model over a generative adversarial network?

A. Faster generation times

B. Greater variety of subjects

C. Text-based image creation

D. Realistic human images

Why is animation more challenging for generative AI?

A. It cannot create enough frames in a reasonable amount of time

B. It's difficult to get the speed of the animation just right

C. Frames have subtle variations that stand out when played in sequence

D. Words cannot capture movement clearly enough for good prompts

What is one method to indicate which term in a prompt should be emphasized?

A. Write in all caps

B. Bold the term

C. Put it inside brackets

D. Repeat it several times

What is one reason a generative AI model might overrepresent certain groups of people when depicting workplace scenes?

A. Training data may favor those groups

B. Models represent the biases of their programmers

C. Models represent the biases of their users

D. Neural networks conceal the reasons

AI Trailhead Questions

More Quizzes