Robots

1. What is the main difference between early industrial robots like the Unimate and modern robots?

• A. Early robots were made of plastic, while new ones are made of metal.
• B. Early robots could only repeat programmed movements, while modern robots can learn.
• C. Early robots had sensors for sight, while modern robots do not.
• D. Early robots were autonomous, while modern robots require constant human control.

2. What is meant by a 'feedback loop' in the process of robot learning?

• A. Watching a video of a human performing the task.
• B. Downloading a skill directly from a computer server.
• C. Performing an action, measuring the outcome, and adjusting behavior based on the result.
• D. Connecting the robot to a giant map of the world.

3. Which technology allows a vacuum robot to create a map of a room and track its position?

• A. Reinforcement learning
• B. Supervised learning
• C. SLAM (Simultaneous Localization and Mapping)
• D. Neural networks

4. What is a major limitation of modern robots regarding 'intelligence'?

• A. They have no concept of physical laws like gravity.
• B. They cannot use electricity to power their movements.
• C. They only know how to function in empty, white rooms.
• D. They understand geometry and data, but do not actually 'know' the meaning of objects.

1. Why do researchers use simulated environments for robots to 'practice' before working in the real world?

• A. Because robots are too dangerous to be handled by humans.
• B. To allow robots to perform millions of trials in a very short amount of time.
• C. Because robots cannot see colors unless they are in a simulation.
• D. To prevent the robot from learning bad habits from humans.

2. What does the concept of 'embodiment' imply about a robot's learning process?

• A. The hardware design directly influences how the software learns to interact with the world.
• B. A robot can only learn if it looks exactly like a human.
• C. The soul of the robot is located in its primary motherboard.
• D. Robots learn faster if they are made of heavier metals.

3. How does the 'fleet' learning model differ from individual biological learning?

• A. Fleets are less efficient than single robots.
• B. Individual humans can instantly share their brain's progress with others.
• C. When one robot learns a new skill, it can share that data with the entire fleet to update all robots simultaneously.
• D. Fleets require humans to manually update every robot individually every time.

4. What is the main benefit of 'transfer learning' for a robotic system?

• A. It allows the robot to copy its code to a new computer.
• B. It enables a robot to apply foundational knowledge from one task to a new, similar task.
• C. It helps the robot avoid learning anything that might be dangerous.
• D. It converts the robot's physical sensors into digital code.