WGU C951 Introduction to AI
Access The Exact Questions for WGU C951 Introduction to AI
💯 100% Pass Rate guaranteed
🗓️ Unlock for 1 Month
Rated 4.8/5 from over 1000+ reviews
- Unlimited Exact Practice Test Questions
- Trusted By 200 Million Students and Professors
What’s Included:
- Unlock Actual Exam Questions and Answers for WGU C951 Introduction to AI on monthly basis
- Well-structured questions covering all topics, accompanied by organized images.
- Learn from mistakes with detailed answer explanations.
- Easy To understand explanations for all students.
Access and unlock Multiple Practice Question for Introduction to AI to help you Pass at ease.
Free WGU C951 Introduction to AI Questions
Two independent events have probabilities P(A) = 0.3 and P(B) = 0.4. What is the probability that both occur?
-
0.12
-
0.35
-
0.70
-
0.10
Explanation
Explanation:
Correct Answer: (A) 0.12 For independent events, the joint probability is the product of the individual probabilities: P(A and B) = P(A) × P(B) = 0.3 × 0.4 = 0.12. Independence means that knowing one event occurred does not change the probability of the other.
Why Other Options are Incorrect:
- B. 0.35 This is the average of the two probabilities, which has no role in computing a joint probability.
- C. 0.70 This is the sum of the probabilities. Adding would be appropriate for mutually exclusive events "A or B," not for "A and B."
- D. 0.10 This does not result from multiplying 0.3 and 0.4.
In k-means clustering, a cluster contains the points (2, 4), (4, 6), and (6, 8). What is the new .centroid after recomputation?
-
(3, 5)
-
(4, 6)
-
(5, 7)
-
(6, 8)
Explanation
Explanation:
Correct Answer: (B) (4, 6) The centroid is the mean of the points in each dimension: x = (2 + 4 + 6) ÷ 3 = 4 and y = (4 + 6 + 8) ÷ 3 = 6. K-means alternates between assigning points to the nearest centroid and recomputing centroids as cluster means until assignments stop changing.
Why Other Options are Incorrect:
- A. (3, 5) This is the midpoint of only the first two points, ignoring the third.
- C. (5, 7) This is the midpoint of the last two points, ignoring the first.
- D. (6, 8) This is simply the last point, not the mean of all three.
What is the key difference between a goal-based agent and a utility-based agent?
-
A goal-based agent has no sensors, while a utility-based agent has many
-
A goal-based agent distinguishes only between goal and non-goal states, while a utility-based agent measures how desirable each state is and can trade off competing objectives C. A goal-based agent can learn, while a utility-based agent cannot
-
A goal-based agent acts without considering the future
Explanation
Explanation:
Correct Answer: (B) A goal-based agent distinguishes only between goal and non-goal states, while a utility-based agent measures how desirable each state is and can trade off competing objectives A goal is a binary notion: a state either satisfies it or it does not. Many real problems involve degrees of success, such as a route that is fast but risky versus slower but safe. A utility function assigns a numerical score to states, allowing the agent to compare outcomes, weigh uncertainty, and resolve conflicts between objectives.
Why Other Options are Incorrect:
- A. A goal-based agent has no sensors, while a utility-based agent has many Both agent types require sensors to perceive their environment, so this does not describe the distinction.
- C. A goal-based agent can learn, while a utility-based agent cannot Learning is a separate dimension. Either type can be built with learning components, and the statement reverses nothing meaningful about their definitions.
- D. A goal-based agent acts without considering the future Goal-based agents specifically reason about the future consequences of actions in order to reach a goal.
Which of the following events is often cited as the formal birth of the field of artificial intelligence?
-
The launch of the first smartphone
-
The 1956 Dartmouth workshop
-
The creation of the World Wide Web
-
The invention of the transistor
Explanation
Explanation
Correct Answer: B) The 1956 Dartmouth workshop
In the summer of 1956, John McCarthy, Marvin Minsky, Claude Shannon, and Nathaniel Rochester organized a workshop at Dartmouth College. McCarthy coined the term "artificial intelligence" in the proposal for this event.
The workshop brought together researchers who went on to lead early AI work in areas such as problem solving, logic, and language.
Why the other answers are wrong:
A) The launch of the first smartphone: Smartphones came decades later.
C) The creation of the World Wide Web: The World Wide Web (1989 to 1991) is unrelated to the founding of AI.
D) The invention of the transistor: The transistor (1947) was crucial for computing in general but did not establish AI as a field.
What is a heuristic in AI search?
-
The total memory used by the search
-
A guaranteed exact solution
-
A random number used to select moves
-
An estimate of the cost from a state to the nearest goal
Explanation
Explanation
Correct Answer: D) An estimate of the cost from a state to the nearest goal
A heuristic function h(n) gives an educated guess of how far a state is from the goal. For route finding, straight-line distance to the destination is a common heuristic.
Informed search algorithms use heuristics to focus on promising states, which can dramatically reduce the number of states explored.
Why the other answers are wrong:
A) The total memory used by the search: A heuristic estimates distance to the goal, not memory use.
B) A guaranteed exact solution: A heuristic is an estimate, not an exact answer.
C) A random number used to select moves: A heuristic is based on problem knowledge, not randomness.
In NLP, what is tokenization?
-
Translating text into another language
-
Breaking text into smaller units such as words or subwords
-
Encrypting text for security
-
Counting the number of documents
Explanation
Explanation
Correct Answer: B) Breaking text into smaller units such as words or subwords
Tokenization is usually the first step in processing text. The sentence "AI is useful" might become the tokens "AI," "is," and "useful."
Later steps, such as counting words or feeding tokens into a model, work on these tokens rather than on raw text.
Why the other answers are wrong:
A) Translating text into another language: Translation is a separate NLP task.
C) Encrypting text for security: Encryption is a security process unrelated to tokenization.
D) Counting the number of documents: Counting documents is not tokenization.
A company deploys an AI hiring tool trained on ten years of past hiring decisions, and the tool begins systematically favoring candidates from one demographic group. What is the most likely cause?
-
The model has learned historical bias present in the training data
-
The model is too simple to learn anything
-
The model has no training data
-
The model was tested on too much data
Explanation
Explanation:
Correct Answer: (A) The model has learned historical bias present in the training data Machine learning models learn patterns from the data they are given, including unfair patterns. If past hiring decisions reflected human bias, the model can reproduce and even amplify that bias while appearing objective. Responsible AI practice includes auditing training data, measuring fairness across groups, maintaining transparency, and keeping humans accountable for high-stakes decisions.
Why Other Options are Incorrect:
- B. The model is too simple to learn anything Learning a biased pattern requires the model to have learned something, so this does not explain the systematic preference.
- C. The model has no training data The scenario states that the tool was trained on ten years of hiring decisions.
- D. The model was tested on too much data. Extensive testing does not cause biased behavior. It is more likely to reveal it.
Which process is used to train neural networks by calculating how much each weight contributed to the error and adjusting the weights to reduce it?
-
Breadth-first search
-
Backpropagation with gradient descent
-
Forward chaining
-
K-means clustering
Explanation
Explanation
Correct Answer: B) Backpropagation with gradient descent
During training, the network makes a prediction and the error (loss) is measured. Backpropagation works backward through the network to calculate how each weight affected the error.
Gradient descent then adjusts each weight a small step in the direction that reduces the error. Repeating this many times gradually improves the model.
Why the other answers are wrong:
A) Breadth-first search: BFS is a search algorithm.
C) Forward chaining: Forward chaining is a rule-based reasoning method.
D) K-means clustering: K-means is an unsupervised clustering algorithm.
A disease affects 1% of patients. A model predicts "no disease" for everyone and achieves 99% accuracy. Why is accuracy misleading here?
-
The model has too many features.
-
Accuracy cannot be calculated for medical data.
-
99% accuracy means the model is nearly perfect.
-
The model misses every patient who actually has the disease.
Explanation
Explanation
Correct Answer: D) The model misses every patient who actually has the disease.
When classes are imbalanced, a model can achieve high accuracy simply by always predicting the majority class. Here, the model never detects a single true case.
Metrics such as recall (sensitivity), precision, and the F1 score give a better picture. For a disease screening model, high recall is especially important so that sick patients are not missed.
Why the other answers are wrong:
A) The model has too many features: The number of features is not the problem; the imbalanced classes are.
B) Accuracy cannot be calculated for medical data: Accuracy can be calculated; it is just misleading in this case.
C) 99% accuracy means the model is nearly perfect: High accuracy here hides a model that is useless for its purpose.
A fraud detection dataset has 99% legitimate transactions and 1% fraudulent ones. A model that labels every transaction "legitimate" achieves 99% accuracy. Why is accuracy a poor metric here?
-
Accuracy cannot be computed on binary data
-
The model never detects any fraud, so a high accuracy hides the fact that recall for the fraud class is 0%
-
Accuracy is only valid for regression problems
-
Legitimate transactions should not be counted
Explanation
Explanation:
Correct Answer: (B) The model never detects any fraud, so a high accuracy hides the fact that recall for the fraud class is 0% With heavily imbalanced classes, a trivial model that always predicts the majority class can score high accuracy while being useless for the task. Metrics such as precision, recall, F1-score, and the confusion matrix reveal how well the model handles the minority class that actually matters.
Why Other Options are Incorrect:
- A. Accuracy cannot be computed on binary data Accuracy is computed easily for binary classification; the issue is interpretation, not computability.
- C. Accuracy is only valid for regression problems The reverse is closer to true: accuracy is a classification metric, and regression uses error measures such as MSE.
- D. Legitimate transactions should not be counted They must be counted; the problem is relying on a metric that is dominated by the majority class.
How to Order
Select Your Exam
Click on your desired exam to open its dedicated page with resources like practice questions, flashcards, and study guides.Choose what to focus on, Your selected exam is saved for quick access Once you log in.
Subscribe
Hit the Subscribe button on the platform. With your subscription, you will enjoy unlimited access to all practice questions and resources for a full 1-month period. After the month has elapsed, you can choose to resubscribe to continue benefiting from our comprehensive exam preparation tools and resources.
Pay and unlock the practice Questions
Once your payment is processed, you’ll immediately unlock access to all practice questions tailored to your selected exam for 1 month .