Knowledge representation: -(Another book matches syllabus though) -Knowledge is a collection of beliefs that an AI agent holds about the world, in other words, refers to the information or understanding that is acquired through learning or experience -Is a complex and multifaceted concept that encompasses many different types of knowledge, including factual knowledge (e.g Paris is capital of France), procedular knowledge (How to ride bike), and conceptual knowledge (eg The concept of justice) -(humans le knowledge use garira hunxa tei bhayera knowledge ma focus gareko bhanni kura not just algorithmic followups) -Knowledge representaion is a scheme to represent diverse factors about the real world in a form that can be used to reason and solve problems -By representing knowledge in a structure and formal way, AI systems can reason, infer and make predictions based on that knowledge leads to more efficient as well as teh ability to learn and adapt to new situations -Particular knowledge representation models allow for more specific, more powerful problem solving mechanisms that operate on them -EUTA SEMANTICS WALA FIGURE TA CHAINXA YAR YETA -A good system for representation of knowledge in a particular domain should possess the folowing properities: 1. Representatioanal adequacy: the ability to represent all of the kinds of knowledge that are needed in that domain 2. Inferential adequacy: The abilty to manipulate the representatinal strucures in a way as to derive new structures corresponding to new knowledge inferred from te world 3. Inferential effieciency: abilty to incorporate into the knowledge structure additional informatino that can be used to focus the attention of the inference mechanisms in the most promising directions 4. Acquisitional efficiency: the ability to acquire new inforamtion easily, the simplest case involves direct insertion, by a person, of new knowledge into the database, ideally the program itself would be able to control knowledge acquistion
-Approahces:
1. Simple relational knowledge
-Simplest way to represent declarative facts is a set of relations of the same sort used in database systems
-(Euta talbe banauni Player name, height weight ra skills ko
-Reason this is simple is that standing alone it provides very weak inferential capabilities but knowledge represented in this form may serve as the input to more powerful inference engines
-Providing support for relatinal knowledge is what database systems are designed to do
Pros:
-Represent complex relationships between entites or objects
-Used in databse to efficiently store and retrieve information
Cons:
-Weak inferential capabititlies
-May not be suitable for representing certain types of knowledge such as rules or procedures
2. Inheritable knoweldge:
-Knowledge is represented as a hierarcy of concepts, with each concept inheriting properties from its parent concept
-Here example of additional baseball knowelge inserted into a structure that is so arranged
-Lines represent attributes, boxed nodes represent objects and values of attributes of the objects, these values can also be viewed as objects with attribute and values and so on
-It may also be called semantic network or collection of rame
-The general atributes are isa, which is being used to show class inclusion, and the attribute instance, which is being used to show class membership
-Now to respond to a query follow the instance lines, or go up the isa hierarchy
Pros:
-Easy to organize and classify information
-Reuse of knowledge, strucutered, modular approahc
Cons:
-May not be suitable for ceratin knowledge such as rules or procedures
3. Inferential:
-Knowledge is repreented as a set of rules or precudres for making inferences or drawing conclusions
-For ex, in a medical diagnosis system, the system can infer a patient's condition based on the symptoms they exhibit
-Yesma ta tei hamro first order logic haru nai bhaye haina ta, jasma hamle resolution with unsatisfiability lagayera inference garna sakinxa
-Example ko lagi feri basketball nai thik xa j lekhey ni bhayo aba ta
Pros:
-Easy resoultion avaiable xa ni ta yesma
-Widely used to learn from data and improve accurarcy over time
-Mathematical backgrounds and theorem suppport
Cons:
-May not be suitable for some relatinoshis such as relationship betwewen entities
-Does not allow uncertaintly, yo ta sablai bhanna milxa hehe
4. Procedural:
-Knowledge is represented as a set of procedures or algorithms for performing specific tasks
-Can be represented in many ways, commonly as code in some programming language such as LISP for doiong something
-The machineuses the knowledge when it exectues the code to perform a taks
-Gets low scores with respect to the proerpties of inferential adequacy because it is very difficult to write a program that can reason about another progrma's behavior
-Comonly used technqiue for representing procedural knowledge in AI is the use of producion rules
-Production rules, particularly ones that are augmented with information on how they are to be use, are more procedular than are the other representatio methods
Pros:
-Good for repeatitive taks and all
Cons:
-Inferential adequacy xaina because difficult ot write
ISSUES:
-Are any attributes of objects so basic that they occur in almost every problem domain? If there are, we need to make sure that htye are handled properly in each of the mechanisms we propose, if such attributes exist, what are they?
-Are there any important relationships that exist among attributes of objects?
-At what level should knowledge be represented? is there a set of primites into which all knowledge can be broken down? Is it helpful to sue such primitives?
-How should set of objets be represnted?
-GIven a large amount of knoweleg stored in a database, how can relevant parts be accessed when they are needed?
Semantic nets:
-is a knowledge base that represents relations between concepts in a network
-In a semantic net, information is represented as a set of nodes connceted to each other by a set of labelled arcs, which represent relationships among nodes
-Main idea is that meaning of a concept comes from the ways in which it is connencted to other concepts
-Example ta tei mathik ko basketball wala lekhdiye bhayo
-The network contians examples of both the isa and isntance relations, as well as som e othe rdomain specifics relations
-Euta hiearhcy bata badheko example dkehauna paryo ni ta hehe
--Logic ra semantics nets ko relationships hai ta:
1. Semantics are a natural way to represnt relationships that would appear as ground istances of binary predicates in predicate logic tyo ta bhaihalyo meaning:
isa(person, mammal) lai kasari garauni ta
team(ssdfs, jsdlfsd) lai kasari garauni ta
2. But he knowledge represetned by predicates of other aritites can also be represnted as semantic nets, for example
man(Marcus) as instance(marcus, man)
3. Three or more place predicaetes can also be converted to binary form by creating one new object represnting the entire predicat e and them introducing binary predicates to descrite the relationsihip to this new object of each of the oringial arguments
score(nepal, srilankda, 1-1) ma aba five ot aobjects hunxan:
game name as a center which conencts with isa to game and otehrs surroudning the gamename
Frames:
-A frame is a collection of attributes usually called slots and associated values and possibly constraints on values that describe some entity in the world
-Sometimes a frame describes an entity in some absoulte sense; sometimes it represents the entity frm a particular point of view
-A single frame alone is rarely useful isntaed, we build frame system out of collections of frames that are connected to each other by virtue of the fact that the value of an atrribute of one frame may be another frame
-The set theory provides a good basis for understanding frame systems, a frame represents either a set or an instance (an element of set)
-Tei aghi mathiko wala lai consider garey bhaihalyo
-Frames add a third dimension to the smeantic nets by allowing nodes to have structures
-Each piece of information about a particular frame is held in a slot, can contain:
1. Facts or data: values
2. Procedures: IF NEEDED: deferred evaluation, iF ADDED: updates linked information
3. Default values: for data and for procedures
4. Other frames or subframes: other frames or subframes
-Because a class represents a set, there are two kinds of attributes that can be associated with it, about the set itself, and there are attriutes that are to be inherited by each element of the set
-Example of lagi euta list banauni plus diagramatically dekhamla k xa ra
Conceptual dependney and scripts:
-CD is a theory of how to represne the kind of knowledge about events that is usually contained in natural language sentences
-Goal is to represent knowledge in a way:
>Facilitates drawing inferences from the sentences
>Is indepdent of the langauge in which the sentences were orignially written
-CD represenation of a sentence is built not out of primitives corresponding to the workds used in the sentences but rahter out of coneptual primites that can be combined to form meanings of words in any particular language
-Over semantic nets is that CD provides both a strucutre and a specific set of primitives, a a particular level of granulity out of which representation of particular pieces of inforamtion can be constructeud
-I gave the man a book can be converted as:
where symbols ko meaning explain gar yo ghokihaal
-Some typical priitives:
ATRANS: transfer of an abstract relationship (give advice)
PTRANS: transfer of the physical location of an object (give book)
MTRANS: transfer of mental informatoin (tell)
SPEAK: production of sounds (speak)
MOVE: movement of a body part by its owner
GRASP:
Scripts:
-A script is a structure that describes a stereotyped sequence of events in a particular context,
-A script consits of set of slots, associated with each slot may be some information about what kinds of values it may contain as well as default value to be used if no other inforamtion is avaiable
-In script as comparison to frame we can make some more precise statements about its structure
-Important concepts of scripts
1. Entry condition: conditions that must in general be satisfied before the events described in the script can occur
2. Result: condition that will be true after the events desribed in the script have occured
5. Track: Specific variation on a more general pattern that is represented by this particular scripts
3. Props: Slots representing the objects that are involved in the events described int he script,
4. Roles: Solts representing the people who are involved in the events desripbed in the script
6. Scnces: actual sequences of events