User talk:Doyoudovoodoo

A HEURISTIC is medium to poorly understood relationship. (Also called "rules of thumb”.) Heuristics are often expressed in the same way as facts or rules (IF … THEN ….), but are given a CONFIDENCE FACTOR (CF). Confidence factors are expressed as a percentage. (Similar to probability)

CF = 0% means a fact or rule cannot happen (an impossible event) CF = 100% means a fact or rule is a certainty. CF = 50% means a fact or rule as a 50-50 chance of being true.

CHAINING

Forward chaining starts with the facts and works through the rules towards a answer. Initial sate to goal state.

Backward chaining assumes that a statement is true and then tries to prove it by checking through the rules and facts. From goal state to subproblems which are solved.

SEARCH TREE CONCEPTS

Terms

Node: A point containing data.

Parent node: The node above the selected node

Child node: The node below a selected node

Root node: The first, highest level node (Level 0). It always unique

Goal node: The nodes at the end of each path of a search tree.

Properties of a search tree

	At level 0 (Highest level) there is a unique node with no parent nodes. 	Each child node has only one parent, but each parent node can have more than one child.

Explicit Definition of a search tree:

Using a diagram to define the tree

Implicit Definition of a search tree:

Using a set of rules or mathematical relationships to generate a tree.

Searching a tree

We can use a Heuristic search to help us search a tree. This uses information and knowledge to help focus the search. This is the common type used in Expert Systems where the questions asked allow the user to search the true using the inherent Heuristics and rules.

Sometimes a tree may not have rules to follow, so a general approach is used. These general approaches are a series of algorithms. Examples are Depth first search, Breadth first search.

Final notes:

	A search tree can be defined as EXPLICIT or IMPLICIT explicit - a diagram is used to define the tree implicit - a rule or mathematical relationship is used to generate the tree 	Depth first searches follow a concept called Progressive deepening. A depth first search is performed to a predetermined level. If a solution is not found then the search continues again using the next depth limit.

	PROBEM REDUCTION

Divide a complex problem into a series of smaller problems. The original problem is solved by solving all the Subproblems.

	Tower of Hanoi

http://www.mathsnet.net/puzzles/hanoi/index.html

Theory Notes on AI part 2

AI Definition

Artificial Intelligence is the study of ideas that enable computers to be intelligent.

What do we define as intelligent.

If I could multiply 5 digit numbers together in a second would I be considered intelligent?

If so a calculator must be a master of AI?

Intelligences form multiple realms – Logical, Arithmetic, Problem solving, Communication, Reasoning and judgment.

Some intelligences are harder for Artificial intelligence to cope with.

Natural language processing - Syntax vs Semantics

Syntax is the structure of a language (Words, Spacing, Characters, etc). Computers can understand this easily – for example a spell checker can find mistakes in documents we miss.

Syntax if written correctly is easily understood.

Semantics is the meaning of language. Computers have trouble here as phrases that structurally look incorrect may have certain meaning.

Eg: Break a leg, Chill out, Don’t have a cow man.

Semantics leads to the idea of natural language processing. The ability to speak and be correctly replied to. Computers with this ability are emerging with the advent in Japan of a help Kiosk that features a computer generated human that responds to conversation. A perfect system is not yet in place. The problem here is not only semantics in language, but the variation in accents, voice tones, volume, pace, etc

Points of breakthrough for AI

	Until the last 70s AI whilst showing signs of development never really showed signs of become equal to humans in intellect. In the 1970s they started beating us at our own games.

	Prospector – An expert system designed to locate ore deposits from raw data Landstat photos (Showing colours of deposits). It found deposits several experts in the field missed.

	5th generation computers – Able to hold conversations with the user and think for themselves

	Deep Blue, a highly parallel 32-node IBM RS/6000 SP supercomputer, beat Gary Kasparov, world champion of chess. Deep Blue did this by calculating hundreds of million of alternative plays for a number of moves ahead.

	An artificial intelligence system, Remote Agent, was given primary control of a spacecraft for the first time. For two days Remote Agent ran on the on-board computer of Deep Space 1, while 60 million miles from earth.

Act: Choose one of the breakthroughs above and spend 10 minutes researching topic. You will be expected to give more details about topic.

Theory of the Mind vs Body

Interactionism – Claims that the mind is a spiritual entity that exists separate to the body. The brain is simply a processing box that acts as a medium for the mind which cannot be physically created.

Materialism – Claims that the mind is part of the physical body. The mind acts in a physical sense and acts in response to messages from the body. This means the mind is a physical entity.

Act: Which of the two theories support the notion of a computer one day being able to achieve human like status and why?

Problem solving.

As humans we are excellent at problem solving. It used in many aspects of life. This is an important area of research in AI.

We solve problems in various ways. We may have a method we follow, an equation to use or apply basic Trial and Error.

Trial and error, while inefficient leads to answer through multiple attempts at a problem. As well as this we use Heuristics (Rules of thumb) to guide us by providing basic rules to follow when using a trial and error system.

AI systems at present tend to be designed to solve problems limited to a particular field or UoD (Universe of discourse) DEPTH and BREDTH

DOYOUDOVOODOO