ARTIFICIAL INTELLIGENCE UNIT-IIT Approaches to Knowledge Representation: There are different Kinds of Knowledge representation Simple Relational Knowledge This type of knowledge uses relational methods to store facts, such as tables or graphs The simplest way to represent declarative facts is a set of relations of the same sort used in the database system. The table below shows a simple way to store facts. Player Height Bats-Throws Hank Aaron 6-0 Right-Right Willie Mays 5-10 Right-Right Babe Ruth 6-2 Left-Left Ted Williams 3 Lefi-Right player_info(*hank aaron’, °6-0", 180,right-right). Fig. 4.4 Simple Relational Knowledge and a sample fact in Prolog * Also, The facts about a set of objects are put systematically in columns. © This representation provides little opportunity for inference Inheritable knowledge: In the inheritable knowledge approach, all data must be stored into a hierarchy of classes. All classes should be arranged in a generalized form or a hierarchal manner In this approach, we apply inheritance property. Elements inherit values fiom other members of a class, This approach contains inheritable knowledge which shows a relation between instance and class, and it is called instance relation. Every individual frame can represent the collection of attributes and its value. In this approach, objects and values are represented in Boxed nodes. We use Arrows which point from objects to their values, Example: Pavani Paladugu, Dept of Computer Science Sir C RReddy College,Eluru.ARTIFICIAL INTELLIGENCE UNIT-IIT Peter 3. Inferential knowledge: © Inferential knowledge approach represents knowledge in the form of formal logics. © This type of knowledge refers to knowledge that can be derived from other ‘knowledge using logical reasoning, such as logic or ontologies. © This approach can be used to derive more facts: © It guaranteed correctness. Example: Let's suppose there are two statements: a Mareus is a man b.— Allmenare mortal Then it represent as a.man(Mareus) b. Wx :man (x) > mortal (x) 4, Procedural knowledge: © Procedural knowledge approach uses sunall programs and codes which describes how to do specific things, and how to proceed. In this approach, one important rule is used which is If-Then rule. In this knowledge, we can use various coding languages such as LISP language and Prolog language. We can easily represent heuristic or domain-specific knowledge using this approach. But it is not necessary that we can represent all cases in this approach. Py mimi Logic Programming is the combination of two words, logic and programming. Logic Programming is a programming paradigm in which the problems are expressed as facts and rules by program statements but within a system of formal logic. Facts Actually, every logic program needs facts to work with so that it can achieve the given goal Facts basically are true statements about the program and data. For example, Delhi is the capital of India, Pavani Paladugu, Dept of Computer Science Sir C RReddy College,Eluru.ARTIFICIAL INTELLIGENCE UNIT-IIT Rules ‘Actually, rules are the constraints which allow us to make conclusions about the problem domain. Rules basically written as logical clauses to express various facts. For example, if we are building any game then all the rules must be defined. Rules are very important to solve any problem in Logie Programming. Rules are basically logical conclusion which can express the facts. ¥ There are several logic programming system today, the most popular of which is PROLOG. The facts that PROLOG programs are represented on Hom Clause for two reasons, 1, Because of a uniform representation, a simple and efficient interpreter can write. 2. The logic of Horn Clause decidable. (Comte) — (Trot Cm) Cur = | Logical Programming > Typical prolog query can be asked as : Query 1 : 2- singer(somu). Output : Yes. Explauation : As our knowledge base coutains the above fact, so output was 'Yes’, otherwise it would have been 'No’. Query 2 : 2- odd_mumber(7), Output : No Explanation : As our knowledge base does not contain the above fact, so output was ‘No! Prolog uses uppercase letters for varinbles and lowercase for constants Clauses are written with the head preceding the body: ": -" is used for left implication, commas separate literals in the body, and a period marks the end of a sentence ‘Syntactic difference between the logic and PROLOG Representation * In Logic programming, Variable are explicitly quantified. In PROLOG, Variables are implicitly quatified. The distinction between variables and constants is made in Pavani Paladugu, Dept of Computer Science Sir C RReddy College,Eluru.ARTIFICIAL INTELLIGENCE UNIT-IIT PROLOG by having all variables begin with upper case letters and all constants begin with lower case letters or numbers In Logic programming, there are explicit symbols for and( , ) and or. In PROLOG, there is an explicit symbol for and(,),but there is none for or. Instead, disjunetion must be represented as a list of altemative statements any one of which may provide the basis for a conclusion. Tn Logic programming, implications of form “p implies Q” are written as pq. In PROLOG, the same implication is written “backward” as q :- p . his form is natural in PROLOG because the interpreter always works backwards from a goal, and this form causes every rule to begin with the component that mest therefore be matched first This first component is called the head of the rule. Consider the following example 1. Logical representation Vx : pet(x) & small (x) + apartmentpet(x) x : cat(x) © dog(x) — pet(x) x: poodle (x) —» dog (x) 4 small (x) poodle (flufiy) Prolog representation apartmentpet (x) : pet(x), small (x), pet (x): cat (x). pet (x): dog(x). dog(x): poodle (x). small (x): poodle(x) poodle (Alufiy). >Bayves Probal ferference: Bayes’ theorem is also known as Bayes’ rule, Bayes! law, or Bayesian reasoning, which determines the probability of an event with uncertain knowledge In probability theory, it relates the conditional probability and marginal probabilities of two random events. Bayes' theorem was named after the British mathematician Thomas Bayes. The Bayesian inference is an application of Bayes! theorem, which is fundamental to Bayesian statistics. It is a way to calculate the value of P(BJA) with the knowledge of P(A/B) where, P(A) is the probability that event A occurs PB) defines the probability that event B occurs. P(A\B) is the probability of the occurrence of event A given that event B has already occurred. P(BIA) can now be read as: Probability of event B occurring given that event A occurred. ‘PIANB) is the probability events A and B will happen together. Pavani Paladugu, Dept of Computer Science Sir C RReddy College,Eluru.ARTIFICIAL INTELLIGENCE UNIT-IIT In probability theory, Bayes” theorem talks about the relation of the conditional probability of event A with known event B: P(ANB)= P(AIB)*P(B)---—-(1) Similarly.the probability of event B with known event A: P(AnB)= P(BIA)*P(A)-—-(2) Now, we have two expressions for P(ANB), since both expressions are equal to P(ANB), we can set them equal to each other: PAIB)-P(B)= PAA) PCA) PAIB)=P(BA)-A)/P(B) P(BIA)= AIB)-B)/P(A) this is Bayes theorem. >DempsterShafer Theory with Example. The Dempster-Shafer Theory (DST) is a mathematical framework used in artificial intelligence to handle uncertainty and combine evidence fiom different sources. It was developed by Arthur P. Dempster and Glenn Shafer. Unlike traditional probability theory, which assigns probabilities to single events, DST allows for the assignment of probabilities to sets of events, making it more flexible in representing uncertainty. Here are some key aspects of DST: Belief Functions: These fimetions represent the degree of belief in a particular event based on available evidence. Plausibility: This measures how plausible an event is, considering all the evidence that does not contradict it Combination Rules: DST provides rules for combining evidence from different sources, even if they are conflicting The frame of discernment and we'll write it as @. @ is set of all possible mutually exclusive hypothesis means at most one has to be true. For example, in a simplified diagnosis problem, consist of the set 0= {All, Flu, Cold, Peu} All: Allergy Fh: Flu Cold: Cold Pheu: Pneumonia The key fimetion we use is a probability density function, which we denotes asm. The function m is defined not just elements of @ but all subsets of it The quantity m(p)measure the amount of belief that currently assigned exactly the set p of hypothesis. © contains n elements, then there are 2° elements of subsets ‘The sum of all belief value is 1 ‘The diagnosis problem We have no information on evidence to choose between four hypotheses. A: Allergy F:Flu Pavani Paladugu, Dept of Computer Science Sir C RReddy College,Eluru.ARTIFICIAL INTELLIGENCE UNIT-IIT ©: Cold P: Pneumonia {©}=1->the doctor have confidence that disease lie @ only. Step 2:Evidence 1(je the patient is suffering from fever {fever,cold,pnew}=0.6) {E.CP}=0.6 { O}=0.4 } ml Step 3:Evidence 2(ie the patient is suffering from runny nose {Allergy, fever, cold }— 0.8) {AFC }-0.8 {0}-0.2 } m2 Step 4: We identify two belief functions m1 & m2,Let X be the set of subset © to which ml assign a nonzero value and Y be the Corresponding set of m2, We define the combination m3 of ml and m2 to be Then we can compute their combination m3 using following table, which we can derive using the numerator of the combination rule: the value of m3 that the combination rule associated with each of them is computed by multiplying the values of ml and m2 associated with the elements from which they were derived. We produce m3 {F.C}=0.48 {AFC} AF.CP}=0.12 m3 {0}=0.08 Step 5:Now, let m4 cortespond to our belief given just the evidence that the problem goes away when the patient goes on a trip: {A}=0.9 {0}=0.1 m4 We can apply the numerator of the combination rule to produce(where @ denotes the empty set) Pavani Paladugu, Dept of Computer Science Sir C RReddy College,Eluru.ARTIFICIAL INTELLIGENCE UNIT-IIT If any Empty set(@)occurs in table the denominator part we calculated =0.54(denominator is 1-9=1-0.54=0.46) {A}=0.288+0.072=0.36/0.46=0.783 {F.C}=0.104 {AF,C}=0.07 {F.CP}=0.026 {}=0.017 Knowledge representation issues Semantic nets. ‘Semantic networks are alternative of predicate logic for knowledge representation. In Semantic networks, we can represent our knowledge in the form of graphical networks. This network consists of nodes representing objects and ares which describe the relationship between those objects. ‘Nodes are sometimes referred to as objects. Nodes are to represent physical objects, concepts, or situation. Ares as links or edges. The links are used to express relationships. It is also known as associative net due to the association of one node with other. ‘Semantic networks are easy to understand and can be easily extended. This representation consist of mainly two types of relations: a) IS-A relation (Inheritance). b) A-Kind-of relation (AKO) a) IS-A relation: TS-A means "is an instance of" and refers to a specific member of a class. A class is related to the mathematical concept of a set in that it refers toa group of objects. Ex: (bat, blue, t-shirt, art) b) A-Kind-of relation: Pavani Paladugu, Dept of Computer Science Sir C RReddy College,Eluru.ARTIFICIAL INTELLIGENCE UNIT-IIT * The AKO relation is used to relate one class to another: * AKO relates generic nodes to generic nodes while the IS-A relates an instance or individual to a generic class. Simple Example of Semantic net Following are some statements which we need to represent in the form of nodes and ares, ‘Statements: 1) Jemyis a cat 2) Jerry is a mammal. 3) Jerry is owned by Jay. 4) Jerry is white colored. 5) All Mammals are animal. 6) Jerry likes cheese. ¢ is-owned is-colored In the above diagram, we have represented the different type of knowledge in the form of nodes and ares. Each object is connected with another object by some relation Constraint Propagation Constraint propagation is a fimdamental technique used in solving Constraint Satisfaction Problems (CSPs). CSPs are problems where you need to find values for a set of variables that satisfy a given set of constraints. Constraint satisfaction problems (CSPs) to systematically reduce the possible values for variables by enforcing constraints. It helps to simplify problems by eliminating inconsistent possibilities before an explicit search. © Define Constraints: Establish rules that must be followed (e.g., in Sudoku, each row must contain unique numbers). Eliminate Invalid Values: Remove values that violate the constraints Iterate Until Stability: Keep applying constraints until no more values can be eliminated Pavani Paladugu, Dept of Computer Science Sir C RReddy College,Eluru.ARTIFICIAL INTELLIGENCE UNIT-IIT Examples of Constraint Propagation 1. Sudoku Solving: If number can only fit in one cell in a row, it umst be placed there. 2. Scheduling Problems: If a person is unavailable at a certain time, that slot is eliminated from their options Map Coloring: If a country must have a different color from its neighbors, invalid color choices are removed. 6 + Variables are regions - WA, NT, Q, NSW, V, SA, T + Domain for each variable — {red, green, blue} + Constraints — Eg. for WA and NT be {(red, green), (red, blue), (green,red), (green,blue), v (bine,red),{bhue,green)} Multiple solutions, ¢.g. — {WA=blue, NT=green, Q-blue, NSW=green, SA-red, T=green, V-blue} Representing knowledge using rules. For the purpose of solving complex problems encountered in AT, we need both a large amount of knowledge and some mechanism for manipulating that knowledge to create solutions to new problems. A variety of ways of representing knowledge (facts) have been exploited in AI programs. Inall variety of knowledge representations , we deal with two kinds of entities. © Facts: Truths in some relevant world. These are the things we want to represent + Representations of facts in some chosen formalism. These are things we will actually be able to manipulate. One way to think of structuring these entities is at two levels : ‘© the knowledge level, at which facts are described, and * the symbol level, at which representations of objects at the knowledge level are defined in terms of symbols that can be manipulated by programs. The facts and representations are linked with two-way mappings. This link is called representation mappings. ‘© The forward representation mapping maps from facts to representations. * The backward representation mapping goes the other way, fiom representations to facts. Pavani Paladugu, Dept of Computer Science Sir C RReddy College,Eluru.ARTIFICIAL INTELLIGENCE UNIT-IIT Reasoning programs ‘internal Representations Engtish generation English Representation Fig. 4.1 Mappings between Facts and Representations Knowledge representation issues Knowledge Representation (KR) is a core aspect of artificial intelligence (AL) that deals with how knowledge is structured, stored, and processed. However, several challenges arise in representing knowledge effectively. 1. Incompleteness + Itis impossible to represent all real-world knowledge + Example: A medical diagnosis system may not have data on rare diseases. 2. Ambiguity + Some words or phrases have multiple meanings, + Example: "Bank" can refer to a financial institution or a riverbank. 3. Inconsistency + Conflicting knowledge can exist within the same system. + Example: One rule says "Ifit rains, then the ground is wet," but another says "IFit is hot, the ground is dry." 4. Scalability + Large knowledge bases require efficient processing, + Example: The internet has vast amounts of knowledge, but retrieving relevant information quickly is challenging. 5. Expressiveness vs. Computabilit + A highly expressive system may be difficult to compute efficiently + Example: First-order logic is expressive but computationally expensive ‘Common Sense Knowledge © Machines struggle with common sense reasoning. + Example: AT may not understand that "Water is wet" or "People eat food.” Dynamic Knowledge + Knowledge changes over time and must be updated. + Example: Scientific discoveries may invalidate previous knowledge. 8. Uncertainty + Notall knowledge is definite; probabilities may be needed Pavani Paladugu, Dept of Computer Science Sir C RReddy College,Eluru.ARTIFICIAL INTELLIGENCE UNIT-IIT + Example: "It will rain tomorrow" is uneertain and requires probabilistic reasoning. Pavani Paladugu, Dept of Computer Science Sir C RReddy College,Eluru.