Introduction to Machine Learning

Which statement best describes Machine Learning within the broader field of computing?

• Subfield of Artificial Intelligence focused on learning

• Programming paradigm for writing deterministic rules

• Independent field unrelated to Artificial Intelligence

• Hardware discipline optimizing processor architectures

The phrase learning from data refers to what core idea in Machine Learning?

• Models rely on human intuition without datasets

• Algorithms memorize fixed programs and routines

• Systems infer patterns from labeled or unlabeled examples

• Developers manually encode every decision rule

In the email spam scenario, what is the task T that the program aims to improve with experience?

• Classifying emails as spam or not spam

• Watching you label emails over time

• Maximizing the fraction correctly classified

• Designing the mail client interface

Which statement best defines supervised learning in machine intelligence?

• Learning by random exploration of environments

• Learning without any target or feedback signals

• Learning with labeled examples guiding outputs

• Learning from rewards with unknown state transitions

How could the house-price scenario be converted into a classification problem?

• Model unsupervised feature extraction

• Estimate average market elasticity

• Predict exact continuous sale price

• Predict sell-above-or-below asking price

Given a face image, estimating the person’s age most naturally requires:

• Regression of a continuous age value

• Clustering into demographic groups

• Classification into tumor categories

• Dimensionality reduction without labels

Which task best exemplifies regression in supervised learning?

• Predicting house sale price as a continuous value

• Classifying emails into spam or not spam

• Assigning documents to topics like sports, tech

• Detecting objects as car, bus, or bicycle

In the housing example, what is the output variable the model predicts?

• Continuous price measured in dollars

• Ordinal quality rating from one to five

• Binary sold or not sold outcome

• Categorical neighborhood label

Which statement correctly distinguishes classification from regression?

• Both produce probabilities but not final labels

• Classification predicts discrete categories, regression predicts continuous values

• Classification predicts continuous values, regression predicts discrete categories

• Both predict continuous values with rounding to cents

Why does the term “supervised” apply to these learning problems?

• Algorithms run under human supervision during inference

• Outputs are restricted to supervised classes only

• Models supervise feature collection in sensors

• Training data includes right answers for each example