Organizer: Microsoft and LinkedIn Learning
Exam Answers Career Essentials in Generative AI by Microsoft and LinkedIn Certification
Q. Your company is developing a smartphone application that identifies plants using the phone’s camera. The company purchases millions of digital images of plants labeled with the name of the species. You use this initial batch of images to train your artificial neural network. What type of machine learning is this?
self-supervised learning
unsupervised machine learning
reinforcement learning
supervised learning
Answer supervised learning
Q. In machine learning, when a data model performs exceptionally well during the training set phase, but lacks the complexity to generate accurate predictions during the test set phase, the model is the data.
competing with
overfitting
reinforcing
underfisting
Answer. Overfitting
Q. You work for a political organization that does sentiment analysis of social media networks. Politicians look to your service to see how people feel about certain difficult topics. Your organization has developed an artificial neural network that can search social media for topis and classify the comments as strongly agree, neutral, and strongly disagree. What type of machine learning are you using?
reinforcement learning unsupervised clustering
unsupervised learning binary classification
variational auto encoding generative Al
supervised learning multiclass classification
Answer. supervised learning multiclass classification
Q. To improve your fitness, you bought a new smartwatch. Your smartwatch uploads your data to a website for free fitness reports. After you purchased your watch, you got a letter from your health insurance company that your insurance rates will increase. The insurance company has arrived at this decision “based on new data. You called the insurance company, and they would not disclose where they purchased your new health data, but you suspect it’s from the smartwatch. What ethical issue might your smartwatch company have violated?
As of right now, they don’t. The system is just compling other authors views on love
They understand love the same way that hurtans do through experience and heartbreak
They could never understand love, but they could experience love by interacting with humans
Answer. As of right now, they don’t. The system is just compling other authors views on love.
Q. You are having some difficulty dealing with a colleague at work. You asked ChatGPT for advice on how to improve your relationship with the coworker. ChatGPT gives you extremely helpful advice You find yourself intuitively thanking it for its help. Given your interaction, do you think that ChatGPT is strong or weak Al?
is neither because CharGPT is a generative Al system which false outside the distinctions in tradictional artificial intelligence
It is strong Al because ChatGPT gave you genuinely helpful advice that’s the same quality as a human’s
It is weak Al because ChatGPT doesn’t understand what it’s saying’s just gathering information that it found online
It is weak Al because ChatGPT is a good example of artificial general intelligence
Answer. It is weak Al because ChatGPT doesn’t understand what it’s saying’s just gathering information that it found online.
Q. When Google was first developing self-driving cars, they famously said that their autonomous car is not a robotics challenge, but a big data challenge. What did they mean by this?
It means that Google is a big data company and shouldn’t be involved in creating self-driving cars.
It means that self-driving cars are not robots, so you shouldn’t use data to figure out how to drive.
It’s easier to design a car to understand when it needs to turn using data than it is to have it turn using robotics.
it’s easier to design a car to automatically turn than it is to have a car know when it needs to turn.
Answer. It’s easier to design a car to understand when it needs to turn using data than it is to have it turn using robotics.
Q. Why would an artificial neural network want to use a cost function?
The cost function makes sure that the artificial neural network does not take up too many processor resources
The cost function automatically apples to each connection for each neuron affar the network Initializes
A cost function helps the network quantily how wrong it was when making a prediction
The cost function allows the network to appreciate the business constraints when making predictions
Answer. A cost function helps the network quantily how wrong it was when making a prediction.
Q. You work for an organization that sells automobiles online. As part of the service, you have an artificial neural network used to predict the likelihood that a car will sell the next hour. At first the network’s predictions were not very accurate, but then it became more accurate but was off by around 5%. How do you think the network will adjust itself to improve its accuracy?
It will reinitialize all the weights of the connections
it will likely adjust the bias
It will increase the size of the test data
it will feed data into the output layer
Answer. it will likely adjust the bias.
Q. You work for a large financial institution that would like to offer immediate approval for loan applications. Your team has identified four predictors about whether someone will be a good loan candidate: income, credit score, employment, and debt. You develop a systern that will look at each predictor independently and then come up with an overall score. What machine learning algorithm are you using?
naive Bayes
linear regression
K-means clustering
K-nearest neighbor
Answer. Naive Bayes.
Q. Your online movie-streaming business wants to create an artificial neural network that can recommend new movies based on what customers have already seen. The team creates a series of XY diagrams of different film genres. Then it puts the film rating along the X-axis and the duration that people watch on the Y-axis. It then makes a recommendation based on how close movies are to each other on the chart. What type of machine learning algorithm is the team using?
Naive Bayes
reinforcement learning
K-nearest neighbor
learning
Answer. K-nearest neighbor
Q. How is machine learning different from a large business application like Microsoft Word?
Machine learning is infinitely more complex than the largest program like microsoft word.
Microsoft Word is one large application while main learning it a set of techniques and algorithms.
One leasing algorithms for many different business applications
Microsoft Word is part of a suite of applications, while machine learning is one big application
Answer. Microsoft Word is one large application while main learning it a set of techniques and algorithms
Q. You ask your ChatGPT system, “What is it like to be in love?” The generative Al system goes on to explain feelings of connection and romance. How do these systems understand such a complex human emotion?
Answer. They understand because they have general artificial intelligence, which allows them to experience emotion
Q. You work for a company that wants to improve spam filtering for mobile email applications. Your data science team gathers one million messages that have been correctly labeled as spam. You then train an artificial neural network to correctly identify these spam messages. After you train the system, one of the product managers asks why you don’t use those same million messages to test the network for accuracy. How should you respond?
It is a good idea to use the same messages, but the machine learning system can test its accuracy.
An artificial neural network does net use test data like other machine learning systems.
That is an efficient way to train the system without having to find another several milion email messages.
If you use the training data, then you’re not testing how well the system will do in the future to identify spam.
Answer. If you use the training data, then you’re not testing how well the system will do in the future to identify spam.
Q. As a prompt engineer, why use this prompt? “Should we add more justices to the US Supreme Court? Give reasons for and against.”?
QDebate-style questions can give you a better understanding of different perspectives.
It is creating a paradow that will potentially harm the system.
You are allowing the system to brainstorm on a topic with you.
It is allowing you to get two answers to one question, which is an efficient way to lower costs.
Answer. Debate-style questions can give you a better understanding of different perspectives.
Q. What does the term model mean in generative Al?
A model is a generative Al that trains another artificial intelligence on a data set.
A model is a data set of ethical issues.
A model is Al mimicking human behavior.
A model is a set of algorithms that have been trained on a data set.
Answer. A model is a set of algorithms that have been trained on a data set.
Q. What does the term “prompt engineering” mean?
This is when you try to get good results from the reasoning engine with short and effective prompts.
it’s a term for the computer scientists that assemble the system or “prompt” the system to life.
It’s when an engineer prompts the system through a regular series of reboots.
It’s a type of search engine that focuses on maintaining the web inder database.
Answer. This is when you try to get good results from the reasoning engine with short and effective prompts.
Q. Why would you use weights on the connections between the nodes in an artificial neural network?
The weights keep the network from feeding backward.
The weights require much less processing power.
The weights allow the neuron to activate more than once.
These weights allow the network to tune itself.
Answer. These weights allow the network to tune itself.
Q. You work for a large ecommerce website. You sell millions of products, so it’s easy for your customers to get overwhelmed. You decide to use an artificial neural network and machine learning to cluster customer purchases into items that are frequently bought together. The hop is that when someone purchases something like toothpaste, the website will recommend something that’s frequently bought with toothpaste, such as facial cleanser. What type of machine learning are you using?
supervised machine learning.
reinforcement learning.
unsupervised machine learning.
generative artificial intelligence.
Answer. unsupervised machine learning.
Q. You just purchased a new smart music player that can add songs if you simply describe the characteristics you’re looking for. To test out the player you say. “Will you add that famous Johnny Cash song that they turned into a movie?” What type of technology will your music player use to find your song?
It is converting your description into a search by using natural language processing.
It is using unsupervised machine learning to make a recommendation based on your music tastes.
is using reinforcement learning to create a personalized playlist for your music player.
It is using generative Al to compose a new song.
Answer. It is converting your description into a search by using natural language processing.
You are going to use machine learning to try and do a better job predicting the weather. To start out, you just want to classify two weather events: “rain” or “not rain What steps would you take to build this system?
Find labeled weather data, create a small training set of that data, and that set aside more data for the test set.
Use a linear regression to show the trend line from “not rain” to “rain”.
Use reinforcement learning to allow the machine to create rewards for itself based weather on how well it predicted the
Input all the labeled weather. data and allow the system to create its own clusters based on what it sees in the data.
Answer. Find labeled weather data, create a small training set of that data, and then set aside more data for the test set.
Q. You prompt ChatGPT with this question: “Should free health care be a universal human right?” The system responds. “This is a very controversial issue and federal and state governments often struggle with the balance of maintaining the healthcare of its citizens, and their human rights.” What is the challenge with this response?
The system is displaying bias towards a Western perspective.
You are asking the system to weigh in on a partisan issue.
This is a global struggle, and ChatGPT limits the response to a federal and state government. ChatGPT has no way to understand the very human struggle of health care.
Answer. This is a global struggle, and ChatGPT limits the response to a federal and state government.
Q. What does it mean for a reasoning engine to have a “hallucination?”
The system presents a confident response that is actually false.
The system doesn’t understand your question, so it refuses to answer.
The reasoning engine runs out of graphic processing units or GPU cycles.
The reasoning engine reaches artificial general intelligence (AG).
Answer. The system presents a confident response that is actually false.
Q. What is the difference between generative Al and discriminative Al?
Discriminative Al is mostly used in government and university work.
Generative Al tends to not work with digital data.
Generative Al creates content while discriminative Al classifies data.
Discriminative Al creates content while generative Al classifies data.
Answer. Generative Al creates content while discriminative Al classifies data.
Q. What is one of the main reasons that marketers might prefer using text-to-image generation services?
The model allows for greater feedback.
They increase productivity.
They consistently produce more creative output. than humans.
Human illustrators seldom follow specific guidelines.
Answer. They increase productivity.
Q. What does it mean when a web search engine is “crawling?”
It’s a way for a reasoning engine to create and maintain a search index.
Crawling is a continuous process to keep the search engine’s index up to date.
It is looking for blased data on the Internet.
it’s when computer scientists visit each web page and rank it according to value.
Answer. Crawling is a continuous process to keep the search engine’s index up to date.
Q. What is a generative adversarial network (GAN)?
when two neural networks work in opposition, with a generator and a discriminator to improve the output generative.
when a discriminator generates output so that a generator can review it and offer adversarial feedback.
when two neural networks work cooperatively to produce the best output.
when two generative Al organizations compete for the same resources.
Answer. when two neural networks work in opposition, with a generator and a discriminator to improve the output generative.
You manage a radiology department in a large hospital. Your hospital has millions of computed tomography (CT) images. You want to create a system where once someone gets a CT scan, the system will immediately check for anomalies. That way it can be sent for review by a senior radiologist. Which generative Al system might work best for this approach?
an adversarial autoencoder (AAA).
a flexible learning encoding X-ray (FLEXO.
a generative autoencoding network (GAN).
a variational autoencoder (VAEs).
Answer. a variational autoencoder (VAEs).
Q. Oxford dictionary defines plagiarism as “the practice of taking someone else’s work or ideas and passing them off as one’s own.” If you ask ChatGPT to describe a sunset, it will give you a response, but these systems have never experienced a sunset. The only way it could respond is by “passing off ideas as its own.” Does that mean that these generative Al systems are plagiarism machines?
No. these systems may be thought of as experiencing events that it hasn’t experienced.
Yes, reciting what others have written about sunsets is plagiarism.
No, these systems are incapable of breaking the law.
It’s unclear, so there needs to be a new measure of authenticity.
Answer. It’s unclear, so there needs to be a new measure of authenticity.
Q. You use a new text-to-image generating service to create a beautiful artistic landscape. You submit your artwork for an Al-generated artistic award and win third place. The news media picks up the story and uses your image in online articles without any compensation or attribution. Did they violate your copyright protection?
No, they didn’t violate your copyright protection, but they did violate the systems.
Yes, but you’ll have to split the copyright proceeds and attribution with the Al system.
Yes, your image and the prompt engineering phrase can be protected by copyright.
No, because currently Al-generated images can’t be protected by copyright.
Answer. No, because currently Al-generated images can’t be protected by copyright.
Q. You’re trying to get better at prompt engineering, so you decided to try a new technique. You say, “Write a 500-word essay on large language models and hallucinations from the perspective of a computer science graduate student at a university.” What technique are you using here?
You are using role playing to get more accurate responses.
You are using a compression technique by limiting the results to 500 words.
You’re brainstorming with the system about large language models and hallucinations.
You are taking an adversarial approach to get both sides of the story.
Answer. You are using role playing to get more accurate responses.
Q. Why is it much riskier for reasoning engines to be wrong than search engines?
Reasoning engines are much more modern, so the technology should be much more accurate.
Reasoning engines usually take a long time to answer, so you want to make sure that it’s right.
Reasoning engines are conversational, so they can be much more convincing.
Since the reasoning engine is ranking the results, it’s difficult to keep track of the answer.
Answer. Reasoning engines are conversational, so they can be much more convincing.
Q What does it mean to run your code in a “notebook?”
A notebook is a container for code that can be run on cloud services.
A notebook is the name that computer scientists use for generative Al output.
A notebook is where a generative Al model keeps all its data.
Generative Al cant work with traditional programming, so it needs to brundle code in a notebook.
Answer. A notebook is a container for code that can be run on cloud services.
Q. You ask your ChatGPT system, “What is it like to be in love?” The generative Al system goes on to explain feelings of connection and romance. How do these systems understand such a complex human emotion?
They missed an opportunity to charge you for accessing the data on your watch.
They were not transparent about how they use or potentially sell your data.
They did not consider that you are trying to use the fitness watch to get in shape.
They shouldn’t sell your data without giving you a percentage of the proceeds.
Answer. They were not transparent about how they use or potentially sell your data.
Q. Your company produces science fiction and fantasy graphic novels. One of your top illustrators has developed a style that is very strongly associated with your brand. Your company decides to create a generative Al model to mimic their illustrations. This new model can create new graphics in their style in seconds. Now the company will have better control over their brand and increase productivity. What is one of the main challenges with this approach?
it is currently illegal in the United States to mimic the style of working illustrators.
Current generative At models are not doing a very good job mimicking creative illustrators.
The generative Al model will always need to be further trained, so it doesn’t save any time.
It will “normalize mediocrity-the graphics will look the same and lack a creative spark.
Answer. It will “normalize mediocrity-the graphics will look the same and lack a creative spark.
Q. You’re an executive for a software development company. Your company develops only one product. You want to include ethical decision-making into your software development, so you ask a senior developer to also serve as the company’s chief Al ethics officer. What would be one of the challenges with this approach?
Software developers are busy people and they need to focus on technical challenges.
Since you have only one product, your serior developer should always be focusing on software development.
With only one product, there aren’t going to be many ethical Al issues, so you should have your developers focus on developing software.
A chief Al ethics officer sets the ethical direction for the entire company and shouldn’t just focus on the product.
Answer. A chief Al ethics officer sets the ethical direction for the entire company and shouldn’t just focus on the product.
Q. A few years ago, you were dating someone who got very upset after the relationship ended. This person became quite enraged and posted misinformation about you on social media. Fortunately, this person’s rage subsided, and you haven’t thought about it for some time. Recently, you were testing out a new artificial intelligence chatbot. You asked the chatbot what it knew about you. The chatbot included a lot of the misinformation that was posted many years earlier. What is one of the ethical issues that this new Al chatbot is overlooking?
The company is not doing a very good job auditing the system for misinformation.
The company shouldn’t be involved at all with your romantic life.
The company is amplifying information that people might find relevant.
The company should do a better job filtering out information that might sound angry.
Answer. The company is not doing a very good job auditing the system for misinformation.
Q. Your large social media company has decided to open source the data and source code for your chatbot. You recently found out that a foreign government has downloaded your code and set up a chatbot to spread propaganda. The chatbot encourages violence against an ethnic minority group What Al ethics violation might your chatbot release have caused?
Your technology assisted a human rights violation.
Your technology is too easy to implement.
Governments should always be able to use your technology for whatever reason they see fit.
There is now a danger of competition from a large well-funded government.
Answer. Your technology assisted a human rights violation.
Q. You are a board member for a large social media company. For years the company has been stumbling through ethical issues. As a result, you’ve convinced other members of the board to create a statement of ethical values. The executives complained that it’s not the board’s responsibility to create an ethical framework for the company. Who do you think is right?
The board is right because creating a high-level ethical value statement is part of their responsibility.
The board is right because the executives should all be working for the board.
The executives are right because they oversee “executing” the values of the company.
The executives are right because this would be a very high level of micromanagement.
Answer. The board is right because creating a high-level ethical value statement is part of their responsibility.
Q. You work for a large website that sells royalty-free images. Your executives want to create a system that uses existing royalty-free images to create a new generative Al system that produces images without paying artist royalties. At a meeting, the manager asks if it’s ethical to use the artist’s own work to eliminate their livelihood. What is a key ethical question here?
Is there a way that the system could generate art without relying on artists?
Should the site serve the artists or the customers purchasing art?
How can you generate the greatest revenue from the service?
Can the system use the art within the current licensing agreement?
Answer. Can the system use the art within the current licensing agreement?
Q. You and a friend purchase smartphones at the same time. After you buy the phones, you both decide to apply for a credit card from the same company. Both of you are approved for the credit card, but your interest rate is 18%, while your friend’s interest rate is just 8%. You both have the same credit score and similar incomes, so you’re can’t figure out why it’s more expensive for you to use your card. What ethical issue might be at play here?
There isn’t much transparency in how the credit card company assigns interest rates.
The company should always assign the highest interest rate to generate more profits.
The system should generate an end-user license, so users accept the higher interest rate.
The system obviously needs more data, so it doesn’t randomly assign interest rates.
Answer. There isn’t much transparency in how the credit card company assigns interest rates.
Q. You are the chief Al ethics officer for one of the largest companies developing generative Al products. You recently found out that your chief executive gave an interview to a large newspaper. The reporter asked how the chief executive felt about the fact that their products might cause mass unemployment. They responded that, “we’re just focused on building great products. As the chief Al ethics officer, what might you say to the chief executive?
When building powerful Al products, the company should focus only on profitability.
Avoid talking to reporters.
If we keep the conversation focused on our great products, we won’t be dragged into an ethical debate.
When building powerful Al products, the company should consider their impact on society.
Answer. When building powerful Al products, the company should consider their impact on society.