keywords:
Bookmark and Share



Front Back
symbolic representation and symbolic reasoning
  two of the key ideas that underpin Symbolic AI
physical symbol system hypothesis (PSSH)
  the core principle of Symbolic AI.
heuristic search,
finding a good solution involves a search through a space of possible solutions
natural language processing
the creation of machines capable of understanding and responding to human languages such as English;
expert problem solving:
for instance, diagnosing medical conditions from complex sets of symptoms;
planning and scheduling tasks
such as airline scheduling or planning the layout of a factory floor;
logical reasoning,
often under conditions of uncertainty.
symbolic system or a physical symbol system.
A collection of expressions
symbol
some entity that exists in the real world and obeys the laws of physics
symbol structures or expressions
an ordered and structured collection of symbols
    designate
things in the world,  objects

symbol expressions stand for or represent other things
processes
that operate on other expressions within the physical symbol system
    interpretation
The task of identifying a process, applying it to an expression and so changing the expression in some way
state
The combination of all the symbol expressions within a physical symbol system
state space,
The set of states that a physical symbol system can be in
graph nodes
arcs  
we can describe the state space of a problem as a graph: a set of nodes each representing a specific state and arcs connecting the nodes and representing processes that can be applied to that state to transform it into another. The state space is simply the set of all possible states and the possible transformations between them.
predicate name
an expression consisting of a predicate name onfor example
arguments
(the symbols inside the brackets, separated by commas)
expert systems,
designed to capture the behaviour of skilled specialists.
production rules
expert knowledge, often as a set of production rules such as:
IF NOT engine-turns-over
AND NOT headlights-are-bright
THEN problem-with-battery
inference
new knowledge is discovered about a situation
antecedent or premise  
(everything between the IF and the THEN)
consequent or conclusion  
(everything after the THEN).
facts and actions
there are situations in which the system may have to decide whether the conclusion of a rule is a new fact or an action to perform.
declarative knowledge
sometimes called know-that knowledge).
procedural knowledge
know-how knowledge
base knowledge   (also known as object knowledge or domain knowledge  
a system’s knowledge about a domain
meta knowledge
knowledge about that knowledge
meta-reasoning.
the system to reason about what it is doing and how it is doing it.
formalisms
representational schemes
Symbolic logic
The core idea of symbolic logic is to develop an algebra for truth ,  
propositional logic (also called propositional calculus
each symbol in a logical expression represents a fact, or   proposition  
rules of inference
Inference is done by applying defined rules of inference
axioms
a logical statement that is assumed to be true
entailed
a logical relation between sentences of a formal language
Predicate logic (also called predicate calculus
Predicate logic sees the world as collections of objects that have properties and relations with other objects; objects are different from each other because they have different properties (such as names). The logic uses symbols to represent objects (such as Neil and Socrates) and predicates (such as mortal and wet )to pick out properties of objects and relations between them  
quantifiers
allow us to say things about whole classes of objects.
universal quantifier
The universal quantifier means ‘for every’ or ‘for all’;
existential quantifier
When we use a universal quantifier, the sentence is true only if it would remain true whatever value we substitute for the variable.
semantic network,
A semantic network represents a set of entities and relationships between them. The entities can be either objects, collections of objects, or concepts.
    definitional networks
    (which define concepts in terms of other concepts)
    assertional networks (
(which describe sets of assertions or propositions, rather like the predicate logic examples given above).
Frames
slots
Frames share many similarities with semantic networks. However, frames are specialised to deal with taxonomies. Each frame represents an object, as in a semantic network, and has a name and a number of slots that contain information about that object. The slots are attribute–value pairs; the attributes are the slot names and the values are the slot fillers. The most common links between slots are termed is_a links and represent inheritance
Scripts
    for processes and interactions
    conceptual primitives ,
A script is made up from a series of events, with each event expressed in conceptual primitives, such as moving from one place to another, building new information from old, or ingesting some substance.
forward chaining
the system is primed with (or swiftly finds) a set of initial facts. If any subset of these facts match the premise of a rule (the If part), that rule is fired and its conclusions are added to the working knowledge. Reasoning finishes when an interesting conclusion is found or there are no more applicable rules to fire
    backward chaining
    the expert system is primed with a goal, a conclusion that it wants to prove or disprove. The system then searches through the knowledge base to find some combination of rules that lead to that conclusion. The premises of those rules then become subgoals and the expert system tries to prove them in turn, perhaps by using other rules or by asking the user questions.
heuristics
well-founded strategies
state space, also called the search space or
problem space.
  The set of all possible states
fitness
   how good a solution is
x of y cards Next > >|