Unlock hundreds more features
Save your Quiz to the Dashboard
View and Export Results
Use AI to Create Quizzes and Analyse Results

OR
Sign inSign in with Facebook
Sign inSign in with Google
Quizzes > Computers & IT

AI General Knowledge Quiz: See Where You Stand

Quick AI knowledge test. Instant results and helpful explanations.

Editorial: Review CompletedCreated By: Tess WilkUpdated Aug 26, 2025
Difficulty: Moderate
Questions: 20
Learning OutcomesStudy Material
Colorful paper art depicting a robot head, gears, and question marks for AI knowledge quiz.

This AI general knowledge quiz helps you check your grasp of core artificial intelligence ideas with 15 multiple-choice questions, and spot gaps fast. Get instant results with concise explanations, then explore the ai ethics quiz, try our information technology quiz, or challenge yourself with a modern technology quiz.

Which task best fits supervised learning?
Predicting house prices from labeled past sales (because labeled inputs-outputs are provided)
Generating text without any human-provided labels
Discovering topics from unlabeled documents
Finding communities in an unlabeled social graph
undefined
Which is a common goal of unsupervised learning?
Clustering customers into groups based on behavior (because no labels are used)
Predicting next week's sales
Classifying emails as spam or not spam
Estimating a policy's expected reward
undefined
What does overfitting mean in machine learning?
The model performs well on training data but poorly on unseen data (because it memorizes noise)
The model is too simple to capture the pattern
The model trains too slowly
The dataset is too large for memory
undefined
What is the main purpose of dropout in neural networks?
To ensure gradients never vanish
To remove bias from labels
To reduce overfitting by randomly dropping units during training (because it prevents co-adaptation)
To speed up inference only
undefined
What does precision measure in binary classification?
TN / (TN + FP)
(TP + TN) / (TP + TN + FP + FN)
TP / (TP + FN)
TP / (TP + FP): how many predicted positives are correct (because it focuses on false positives)
undefined
In reinforcement learning, what is the reward?
A label provided for each state-action pair by a human
The learned value function
A scalar signal indicating immediate feedback for an action (because it drives policy improvement)
The set of all possible actions
undefined
What is L2 regularization (weight decay) primarily doing?
Penalizing large weights by adding their squared magnitude to the loss (because it discourages overfitting)
Increasing the learning rate over time
Randomly zeroing hidden units during training
Penalizing the sum of absolute values of weights
undefined
What does the bias-variance trade-off describe?
Choosing between precision and recall
A trade-off between CPU and GPU usage
Selecting a batch size
Balancing underfitting and overfitting as model complexity changes (because bias decreases and variance increases)
undefined
What is the F1 score?
AUC of the ROC curve
Geometric mean of accuracy and recall
Harmonic mean of precision and recall (because it balances both)
Arithmetic mean of precision and recall
undefined
In a confusion matrix, what is a true positive (TP)?
A positive example predicted as positive (because both label and prediction are positive)
A negative example predicted as negative
A positive example predicted as negative
A negative example predicted as positive
undefined
What does ROC-AUC primarily evaluate?
Model complexity
Calibration of predicted probabilities
Ranking ability across thresholds (because it measures the trade-off of TPR vs FPR)
Training speed of a model
undefined
Which mechanism helps LSTMs mitigate vanishing gradients compared to vanilla RNNs?
Batch normalization
Gating with cell state that preserves long-range information (because gates control flow of gradients)
Max pooling
Hard threshold activation
undefined
What do transformers use self-attention for?
To compute pairwise interactions among tokens to weigh relevant context (because attention scores relate all positions)
To compress sequences via pooling only
To enforce strictly local context windows
To replace gradient descent
undefined
What is the vanishing gradient problem?
Gradients growing without bound
Gradients becoming too small to update early layers in deep networks (because of repeated multiplications)
Loss going to infinity
Weights turning negative
undefined
What is the core idea of an SVM for linearly separable data?
Use k-nearest neighbors
Find the hyperplane that maximizes the margin between classes (because larger margin generalizes better)
Cluster the data into k groups
Minimize training error only
undefined
The Bellman equation expresses a value as immediate reward plus discounted expected value of the next state.
True
False
undefined
An epsilon-greedy policy chooses the least visited action deterministically at every step.
True
False
undefined
Random forests are prone to overfitting because all trees are trained on the same full dataset without randomness.
True
False
undefined
The kernel trick in SVM requires explicitly computing coordinates in the high-dimensional feature space.
True
False
undefined
k-nearest neighbors requires a gradient descent training phase to learn parameters.
True
False
undefined
0

Learning Outcomes

  1. Analyze foundational AI concepts and terminologies
  2. Evaluate the strengths and limitations of common AI models
  3. Identify key use cases across AI domains
  4. Differentiate between machine learning and deep learning techniques
  5. Apply problem-solving skills to AI scenario questions
  6. Demonstrate understanding of ethical considerations in AI

Cheat Sheet

  1. Understand Key AI Terminologies - Dive into the exciting world of AI by getting comfortable with terms like Machine Learning, Deep Learning, Neural Networks, and Natural Language Processing. These buzzwords are the building blocks of AI magic and help you talk the talk when exploring cool projects.
  2. Differentiate Between Machine Learning and Deep Learning - Imagine Machine Learning as a clever apprentice that learns from examples, while Deep Learning is like a master chef layering flavors for complex dishes. Grasping this difference unlocks a clearer view of how AI tackles tasks from image recognition to voice assistants.
  3. Explore Common AI Models and Their Applications - AI models come in all shapes and sizes - supervised, unsupervised, reinforcement, and more - each with a special superpower. By exploring these techniques, you'll discover why some models excel at prediction, while others shine in decision-making or pattern discovery.
  4. Recognize the Strengths and Limitations of AI Models - Every superhero has strengths and weaknesses, and AI models are no different. They can crunch massive data sets like a champ but sometimes stumble on common-sense reasoning or bias, so knowing their limits keeps your expectations realistic.
  5. Identify Key AI Use Cases Across Domains - From spotting diseases in medical scans to catching fraud in banking systems and powering self-driving cars, AI is everywhere you look. Exploring these real-life case studies shows how versatile and impactful AI can be across industries.
  6. Understand the Basics of Neural Networks - Neural networks mimic the brain's neuron connections, allowing computers to learn and recognize patterns in data. Understanding these layers of virtual neurons is key to building smarter AI systems.
  7. Learn About Natural Language Processing (NLP) - NLP teaches machines to chat, translate, and interpret human language, making chatbots and translation apps possible. It's like giving computers the gift of gab, bridging the gap between binary and everyday speech.
  8. Explore the Concept of Generative AI - Generative AI models, like GPT, are the digital artists of the AI world, crafting brand-new text, images, or music from scratch. Learning how they remix and recreate content reveals AI's creative potential.
  9. Understand the Importance of Ethical Considerations in AI - Designing AI with fairness, transparency, and accountability isn't just a nice-to-have - it's essential. Ethical AI frameworks help prevent bias and build trust, ensuring technology benefits everyone.
  10. Familiarize Yourself with the Turing Test - The Turing Test is the OG challenge asking whether a machine's conversation feels as human as your best friend. Exploring this concept sparks debates about intelligence, consciousness, and what it means to be "smart."
Make a Quiz (FREE) Copy & Edit This Quiz Create a Quiz with AI

Technology Quizzes

Powered by: Quiz Maker