Artificial Intelligence Flashcards
What is the definition of Artificial Intelligence?
Ability to perform tasks that normally require human intelligence (replacing a human activity/thought with a machine)
A.I Graph Search - Difference between Breadth First Search (BFS) vs. Depth First Search (DFS)
Breadth First Search (BFS) - takes a long time
Depth First Search (DFS)
* putting a ton of obstacles (cup shape blocking it), since it’s greedy, it’ll get trapped
* putting the goal in the middle of the map, it might not find it very quickly
What are the three examples of A.I Mike talks about? (name a few)
Potential Fields, Graph Search, Expert System, Support Vector Machines, State of the Art
What is A.I Expert System?
take any human activity (any job) and create a survey paper and try to create these questions about what the profession would do.
ex. through if else statements
What is Genetic Algorithm (GA)?
- inspired by natural selection
- whatever initial parameter is, something will change at random
- control basically until it works, trial and error
- if system does better, you can keep the state
What are Support Vector Machines? (SVM)
Solves single classification or multi-classification problems
VERY dependent on user configuration and Kernel Selection
What is Kernel Selection?
Kernel Selection is the assumption about how data will group
What is State of the Art?
It ties in with Deep learning (based on artificial neural networks with representation learning)
ex. Pic with mom/son
Who is mom? Son? What are they doing?
How to train robots to grasp objects from a bin?
◦ Share learned information with other robots
◦ ‘Rewards’ and ‘Penalties’ exist after every attempt
◦ …some constraints exist to simplify the problem
What are some issues with Machine Learning?
If a condition changes, an entire new network
needs to be trained
Large datasets required
How to train a system with limited data?
What are decision trees?
Decision trees for a task can characterize the task as:
1. wide or narrow, and
2. shallow or deep
More available options make a task wider.
More subsequent, dependent options make a
task deeper.
What are the two kinds of reasoning?
Algorithmic - Creating a set of ordered steps and doing them in a particular order to solve a problem in a way that could be repeated by others
Heuristic - a mental shortcut commonly used to simplify problems and avoid cognitive overload (NO RECIPE TO SOLVE PROBLEM)
Who created algorithmic reasoning?
Muhammad ibn Musa al-Khwarizmi
What’s the goal of A.I?
To make machines that think like people
1) Formalize the way we think
2) Whatever can be formalized can be mechanized
Who created syllogisms?
Aristotle
