Lecture 1 & 2: Introduction Flashcards

1
Q

What is cross-situational learning? Give an example.

A

Learning based on situations or experiences. An example would be how associating a word with an image can have different results. I could associate the word happiness with the beach, but another person could associate it to a family member.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
2
Q

Explain the difference between fast-mapping and mutual exclusivity.

A

Fast-mapping is to show and then identify the object. Mutual exclusivity is when one object is known and one isn’t, the unknown can be identified through a process of elimination.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
3
Q

Explain the iCub Modi experiment.

A

6 experiments were carried out against infants in robots. It involved examining the effect of changing the posture, changing the task, and task interference.

Posture change is changing the position of the subject (moving the robot). Task change is changing the location of the object, so it has no associated location. Task interference is changing the object’s location while identifying the object only.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
4
Q

What new predictions were made in the iCub Modi experiment?

A

Infants have a stronger memory for objects associated with a specific posture. They have better association than robots.

First, they predicted that the effects of posture on object naming would be stronger in robots than in infants.

Second, they predicted that the effects of posture on object naming would be stronger for objects that are more difficult to identify.

Third, they predicted that the effects of posture on object naming would be stronger for objects that are more important to the robot or infant.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
5
Q

What is middleware?

A

The operating system of a robot that integrates camera and sensors.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
6
Q

Explain the difference between cognitive robotics, artificial cognitive systems, and intelligent robots.

A

Cognitive robotics is the combination of methods from AI and cognitive and biological sciences to design intelligent robots.

Artificial cognitive systems are the modelling of the agents. It is the theoretical aspect of simulated/physical agents.

Intelligent robotics is the engineering approach to the design of robots.

KEYWORDS:
- Cognitive robotics -> AI + cognitive science + biological science
- Artificial cognitive systems -> theoretical approach
- Intelligent robotics -> engineering approach

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q

What are the cognitive robotics approaches? Explain each one in one sentence.

A

Developmental robotics is the dynamic process of developing cognitive abilities.

Evolutionary robotics consists of evolutionary computation algorithms.

Swarm robotics is the application of swarm intelligence to portray collective behaviour.

Soft robotics uses soft materials for a more comfortable robot experience.

Neurorobotics is the use of computational neuroscience.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
8
Q

Order the following in chronological order: iCub, Shakey, Octopus, Tortoises, Darwin

A

History: Tortoises (1950) -> Shakey (1966) -> Darwin (1992) -> CB2, iCub (2000s) -> Octopus (2020)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
9
Q

What is cognition? (6 key attributes)

A

Cognition is the process by which an AUTONOMOUS system PERCEIVES its environment, LEARNS from experience, ANTICIPATES the outcome, ACTS to pursue goals, and ADAPTS to changing circumstances.

The 6 attributes: autonomy, action, perception, anticipate, adapt, and learn

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
10
Q

What are the levels of Marr’s levels of abstraction?

A
  1. Computational / Theory level
  2. Representaion / Algorithmic level
  3. Implementation
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
11
Q

Give a robot application example through Marr’s abstraction hierarchy.

A

The example is a robot learning words.

The first level is the theoretical level and encompasses the phenomena to be represented. The phenomena is the process of learning words.

The second level is the representation of the process. The input to the process is a categorical or visual representation. The output is speech through audio or pointing through hands.

The third level is the implementation. It is implemented through a neural network.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
12
Q

Explain one embodied cognition theory.

A

Wilson’s theory explains that the body plays a significant role in the agent’s cognitive processing.

Pfeifer’s embodied intelligence says that cognition can be developed through sensorimotor coordination.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
13
Q

Match each historian with his theory:
Theories:
1. The body plays a significant role in the agent’s cognitive processing.
2. Cognition can be developed through sensorimotor coordination.
3. The theory of grounded cognition.

Historians:
A. Pfeifer
B. Barsalov
C. Wilson

A

1 -> C
2 -> A
3 -> B

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
14
Q

What is Leveneseuqe and Reiter’s cognitive robotics manifesto?

A

Robotic agents require reasoning that leads to a decision on how to act. It is important in the field of AI & knowledge-based systems.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
15
Q

Explain two robot examples implementing synthetic robotics methodologies.

A

*Synthetic methodologies are based on recreating agents with a brain-inspired system

The first example is Walter’s tortoise, which had simple neural circuits.

The second example is Braitenberg’s vehicles, which described a series of theoretical agents with the ability to show fear and love.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
16
Q

What is Brooks contribution to the robotics field?

A

He contributed to behaviour-based robotics by finding that intelligent behaviours can be achieved by reactive architectures.