Problem 1 Flashcards
When is machine learning most useful ?
Environments with a high degree of complexity
–> because we live in a world where the amount of complexity increases exponentially
Why is machine learning so helpful and useful ?
Its accuracy in Prediction, meaning anticipating what will happen in the future
–> take info one has to generate info one previously didnt have
Judgment
Refers to the ability to make considered decisions
–> understanding the impact different actions will have on outcomes in light of predictions
Why is machine learning so efficient and often times more efficient than human learning ?
Because it
a) analyzes thousands of human-to-human actions
b) incorporates the feedback on actions + outcomes to develop a more accurate prediction + new strategies
Cognitive science
Refers to the study of the mind + its processes
–> one studies intelligence + behavior, focusing on how NS represent, process + transform info
The mind cannot be understood if it is only studied at a single level.
Name its 3 levels of analysis.
- Computational level
- Representational + alghorithmic level
- Hardware implementation level (Physical level)
Computational level
Specifies the goals of a process + the logic behind the manner with which it is executed
e.g.: What does the system do ?
Representational + algorithmic level
Adresses how the process can be executed
e.g.: what steps does the system go through
Hardware implementation level
Adresses how algorithm + representation may be physically realized
e.g.: In what ways are the steps the system goes through implemented
Name a few of the different methods that are used to study cognitive science.
- Behavioral experiments
- Brain imaging
- Computitonal modeling
- -> mathematically formal representation of the problem
Artifical intelligence (AI)
Refers to technology designed to perform activities that normally require human intelligence
–> emulates complex human behavior that is capable of
a) reasoning
b) learning
c) acting upon an environment autonomously
Turing test
Refers to a method for judging the intelligence of machines
–> to pass it, a computer program must impersonate a human real-time written conversation sufficiently enough, so the judge can’t reliably tell whether it was a human or machine that wrote it
Name the most important branches of AI.
- Machine learning
- Natural language processing
- -> how computers process human natural languages
Algorithms
Refer to sets of unambiguous instructions that a mechanical computer can execute
–> capable of learning from data as they can better themselves by learning new strategies that worked in the past
There are 4 approaches to AI.
Name them.
- Symbolism
- -> formal logic - Bayesian inference
- -> Bayes theory is used - Nearest neighbor
- Artificial networks
How has psychology contributed to the fields of AI ?
Hebbs theory for how neurons learn through the strengthening of connections between them formed the basis for “Mark 1 perceptron”
–> first mashing that learn on its own
What is the aim of Cognitive science ?
- To describe different kinds of problem solving and learning
- Learning how the mind carries out these operations
Computational representational understating of the mind
CRUM
Suggests that thinking results from the representational structures in the mind + computational procedures that operate on those mental representations
–> in order to produce thought + action
Ai predictions can be divided into 4 types.
Name them
- Timeline + outcome predictions
- -> when we will achieve AI milestones - Scenarios
- -> conditional predictions; if certain conditions are met, certain outcome will follow - Plans
- -> when deciding to implement a certain plan, they’ll be successful - Issues + meta-statements
Are there any disadvantages to implementing AI ?
- Several legal + ethical issues
- Displaces human workers
- Have the potential to repair or improve general cognitive abilities by making humans into cyborgs
- Unknown outcomes associated with technology that approaches human general intelligence
