L1_Basics Flashcards
What is Cognition?
Derived from cognoscere (latin): to recognize
• Refers to processing of information
e.g. attention, memory, learning, problem solving, decision making, language .
• Cognitive processes can be
natural or artificial, conscious or unconscious
• Cognition is studied by
linguistics, neuroscience, philosophy, psychology
AND computer science (artificial intelligence, machine learning)
Artificial Intelligence (AI)
”the science and engineering of making intelligent machines” [McCarthy et al., 1955]
Machine Learning (ML)
a branch of AI that uses statistical techniques to learn from data
- Prediction
- Visualization/Interpretation
- Compression
Discriminant function
The learning machine learns the best possible discriminant function between the classes.
Supervised learning:
Given labeled training examples, learn a mapping from inputs (features) to output (label)
Generate a function that maps input data into output data
Classification and regression problems Labels for training
Unsupervised learning:
Find structure in data e.g. group similar objects together
Allow partitioning of data in meaningful parts No labels for training
Reinforcement Learning:
learn actions in an environment which maximize some notion of reward
e.g. robots learning behaviour
Summary of Cognitive processes:
Perception, recognition of and inference on semantic concepts
Summary of Cybernetics
simple (hardwired) models of cognition
inspired by biological organisms
modeled motion detection, navigation, … → but what about higher cognitive functions?
Summary of Artificial intelligence
adopted ideas from Cybernetics focused on (biologically inspired) models of higher cognition Old AI: rule based systems New AI (machine learning): learns rules from data
Braitenberg Vehicles:
What happens at Fear (+)?
The vehicle drives away from light. It is afraid of light.
Braitenberg Vehicles:
What happens at Love (-)?
The vehicle drives towards light sources and stops in front of them. It loves the light.
