1What are real-world examples of Machine Learning mentioned in the text?
Image labelling (Grumpy Cat), Game opponents (Go/Checkers), Self-driving cars, and Recommendation systems (Netflix/Amazon).
2What is the key requirement for Machine Learning to work in the examples provided?
Big Data (massive amounts of information like photos or ratings).
3How did Arthur Samuel define Machine Learning in 1959?
Giving computers the ability to learn without being explicitly programmed for every single step.
4What is the difference between "Old School" programming and Machine Learning for a game like Checkers?
Old School requires coding every possible move; Machine Learning involves showing the computer thousands of games so it learns the strategy itself.
5In Tom Mitchell's formal definition of machine learning, what do T, E, and P stand for?
Task (T), Experience (E), and Performance (P).
6How does Supervised Learning work?
You give the computer inputs AND correct answers (labels), and it creates a rule to label new data correctly.
7How does Unsupervised Learning work?
You give the computer data without labels, and it must find hidden patterns or groups on its own.
8How does Reinforcement Learning work?
The computer learns by trial and error using a score system (rewards for success, punishments for failure).
9What is the main strength of Traditional Statistics compared to Machine Learning?
Statistics is great for understanding why relationships exist (relationships), while Machine Learning is better for prediction using messy data.
10Why is Machine Learning sometimes called a "Black Box"?
You feed data in and get an answer out, but you might not know exactly how the computer figured it out.
11What does fMRI measure?
It scans the brain to see where blood is flowing, which indicates brain activity.
12What was the "Old Way" of analyzing fMRI data (often called "Blobology")?
Looking for large "blobs" of activity to see where the brain is working, rather than analyzing detailed patterns.
13What is the shift in modern neuroscience regarding brain mapping?
Moving from asking "Does this area light up?" to asking "What is the detailed pattern inside that area?"
14What is a Support Vector Machine (SVM)?
A machine learning tool that tries to draw a line (or a wall) between two groups of data to classify them.
15Why is an SVM useful for analyzing brain cells compared to averaging?
It can distinguish between different patterns (e.g., Cat vs. Dog) even if the average activity level in the area is the same.
16In the experiment with tilted lines, what was the computer able to do?
It could decode which lines a person was looking at (or even imagining) based on their brain activity.
17What biological structure makes orientation decoding possible?
Neurons that prefer specific angles are arranged in "pinwheel" patterns on the brain's surface.
18How did scientists use Machine Learning to communicate with "vegetative" coma patients?
They asked patients to imagine playing tennis (Task A) or walking (Task B) to answer "Yes" or "No," and the computer decoded the distinct brain patterns.
19What did the Dream Decoding experiment prove?
That computers can predict the visual content of dreams (e.g., seeing a person) by recognizing brain patterns during sleep.
20What is Representational Similarity Analysis (RSA)?
A method that measures the "distance" or similarity between two brain patterns (e.g., "How similar is A to B?") rather than just classifying them.
21What major distinction did RSA reveal in how the brain organizes visual information?
The brain separates "Animate" objects (living things like facesodies) from "Inanimate" objects (tools/places).
22How can RSA be used for Hypothesis Testing?
By building computer models based on different theories (e.g., "Brain cares about Aliveness") and seeing which model matches the real brain data best.
23What did scientists discover about individual differences using RSA?
Each person has a unique "fingerprint" of brain similarity that correlates with how they personally perceive the world.
24Which method is considered a "universal language" to compare brain data, behavior, and computer models?
RSA (Representational Similarity Analysis).