END TERM EXAMINATION [MAY-JUNE 2018] SIXTH SEMESTER [BCA] ARTIFICIAL INTELLIGENCE [BCA-312] qime:3 brs. _ MM Note: Attempt any five questions including Q. no, 1 which is compulsory. Select one question from ‘each unit. Q.1. Write short notes (Any Five): (a) Heuristic (6) ‘Ans. Heuristic means (an approach to problem solving, learning, or discovery) that employs a practical method not guaranteed to be optimal or perfect, not following or derived from any theory. Heuristic is a technique designed for solving a problem more quickly when classic methods are too slow, or for finding an approximate solution when classic methods fail tofind any exact solution. ‘The objective of a heuristic is to produce a solution in a reasonable time frame that problem at hand. This solution may not be the best of all is good enough for solving the exact solution. But it is the solutions to this problem, or it may simply approximate the still valuable because finding it does not require a prohibitively long time. r they may be used in conjunction Heuristics may produce results by themselves, be used to with optimization algorithms to improve their efficiency (e.g., they may generate good seed values). Q.1. (b) Task domain (5) ‘Ans, The area of human intellectual endeavour to be captured in an expert system is sometimes called the task domain. Task refers to some goal-oriented, problem-solving activity, Domain refers to the area within which the task is being performed. Some of the typical tasks are diagnosis, planning, scheduling, configuration and design. For example, a program capable of conversing about the weather would be a knowledge- based system, even if that program did not have any expertise in meteorology, but an expert system must be able to perform weather forecasting. * Q.1. (ce) Expert System Ans, Building an expert system is know Practitioners are called knowledge engineers. tbensure to make sure that the computer has all the knowledge needed to solve 0 problem. The knowledge engineer must choose one or more forms in which to represent the paitited know! i knowledge representation ledge ice.. +t choose one or more kno’ a eda emantic nets, frames, (6) n as knowledge engineering and its Itis the job of the knowledge engineer 7 . Bere (A number of knowledge representing schemes like 8 cate logic and rule based systems). 7 &1.@ Hill Climbing 5 ds based on hill climbing get their Ans. C woe he as cet ot a jon. At each point in the search to the top of the hill (the goal) from the way the nodes are selected for expansi be available Sucee; ‘ck 8 select ssor node that appears to lead most quickly e a With wey 2 exploration. This method requires hat some information h to evaluate and order the most promising cbotces: 1a. Scanned with CamScanner22-2018, Sixth Somostor, Artificial Intelligence wit 4 3 B16 Y1D Search with hill climbing. sing, child is selected noice aluated 1. promising is then Hill climbing is like depth-first scarching where the most promi for expansion. When the children have been generated, alternati using some type of heuristic function. The path that appears mos! theses and no further reference to the parent or other children is retained. "This process continues from node to node with previously expanded nodes being discarded. A typical path is illustrated in figure where the numbers by a node correspond to the computed estimates of the goal distanee for alternative paths. ‘The problems encountered with hill elimbing can be avoided using @ best-first search approach. 6) Q. 1. (e) Role Learning ‘Ans, Weall learn new knowledge through different methods, depending on the type of material to be learned, the amount of relevant knowledge we already possess, and environment in which the learning takes place. Five different learning methods under this taxonomy are: + Memorization (Rote learning) + Direct instruction (by being told) + Analogy * Induction + Deduction Q.1.(f) Robotic Architecture : @) ‘Ans, Roboties deals with the design, construction, operation, and use of robots, #° i i well as computer systems for their control, sensory feedback, and information processing: Robotics is a branch of engineering that involves the conception, design, manufactur and operation of robots. This field overlaps with electronics, computer science, artifici#! intelligence, mechatronics, nanotechnology and bioengineering. Architecture * Power source + Actuation ( Blectric motors, Linear actuators, Series elastic actuators, Ai", muscles, Muscle wire, Electroactivé polymers, Piezo motors, Elastic nanotube + Sensing (Touch, Vision, Other,) ° ; « Manipulation( Mechanical grippers, Vacuum grippers, General purpose ecto® Scanned with CamScannerLP. University-[BCA]-Akash Books 2018-3 * Locomotion } (a) Rolling robots, (Two-wheeled bala: Spherical orb robots, Six-whecled robots, Tracked robots) (b) Walking applied to Tobots( ZMP technique, Hopping, (© Other methods of locomotion ( Flying, Snaking, Skati Sailing) ncing robots, One-wheeled balancing robots, Dynamic balancing) ing, Climbing, Swimming, Environmental interaction and navigation Human-rebot interaction (Spéech recognition, Robotic voice, Gestures, Facial expression, Artificial emotions, Personality, Social Intelligence) * Control ( Autonomy levels) Q.1. (g) DFS (Depth First Search) (5) Ans. Depth-frst search of a tre, Depth- first searches are performed by diving downward into a tree as quickly as Possible. It does this by always generating a child node from the most: recently expanded node, then generating that child’s children, and so on until a goal is found or some cutoff. depth point d is reached. Ifa goal is not found when a leaf node is reached or at the cutoff Point, the program backtracks to the most recently expanded node and generate another ofits children. This Process continues until a goal is found or failure occurs. An example ofa depth-first search is illustrated in figure. DEPTH-FIRST SEARCH 1. Place the starting node s on the queue. Ifthe queue is empty, return failure and stop. Ifthe first element on the queue is a goal node g, return success and stop, otherwi a expand the first element, and place the children-at the front of the queue (in any order) . Return to step 2. ex = a ONIRY Deseribe it. (6.5) 2-2. (a) What is the importance of Artificial pate ence +s ff eae a We are a privileged generation to live in this era fu Not eae weir’ ents, Gone are the days when almost everything was genera ‘ana varios Ree hg ta where a lot‘of work is taken over by machines, sof een “atic procaasea ee this regard, artificial intelligence has a spe Scanned with CamScannerSixth Semester, Artificial Intelligence ntelligence or Al is ]™ ines 4-2018 advancement made today. Artificial i nothing but the science of computers and machines “eveloping intelligence like humans, In this technology, 7 i ble to do some of the simple to complex stu 's need to do on a regular basis. As the Af sed on a day to day basis in our daily life, it lives have also become advanced systems are u: js not wrong to say that our with the use of this technology. elp for humans: The Al systems are e! ; efforts sane reas, In order to perform various activities in the industry, any of them are using artificial intelligence to create machine slaves that perform various activities on a regular basis. The artificial intelligence applications help to get the work done faster and with accurate results. Error free and efficient worlds are the main motives behind artificial intelligence. In the recent years, many sectors have started using AI technology to reduce human efforts, and also to get efficient and faster results. What important role does artificial intelligence play today? Widely used in banking and financial systems: It is a known fact that banks have a numerous activities on a day to day basis that need to be done accurately. Most of the activities take a lot of time and efforts from the employees and at times there is also a chance of a human error in these activities so to speak. Some of the works that banks and financial institutions handle are investing money in stocks, financial operations, managing various properties, etc. With the use of AI system in this process, the institutions are able to achieve efficient results in a quick turnaround time. The strategic implementation of artificial implementation in the bank helps them to focus on every customer, and provide them quick resolution. The customers are happy at all times, as they can get quick service for all ~*~ a the banking and financial needs. An important feature of medical science: The Artificial intelligence has completely changed the way medical science was perceived just a few years ago, There are numerous areas in medical science where Al is used to achieve incredible value, With the help of AI, the medical science was able to create a virtual personal health care assistant. These are used for the research and analysis purposes. There are also many efficient healthcare bots introduced in the medical field to provide constant health support to patients. The bots are effective in answering the frequent questions, and also to schedule appointments in quick succession. . ___ Aperfect addition to heavy industries: Artificial A intelligence is used in the production unit in most big ape teen akon paniee Al yntan is used to give a specific the This aunt move objets from one place to another, * application is also used in the management of most Gere to get their tasks efficiently done on time, It is ie cae me ofan employe, the crucial data ial decisi ered and can be easily extracted at the crucial decision making time, Heavy induster ab the on the A system because thy pee a and have the ; : cir tasks done on time potential to put inaccurate data in their system, fficient enough to reduce human Scanned with CamScannerLP. University-[BCA]-Akash Books 2018-5 Efficient use in Air transport: Air transport is one of the most systematic transportation systems, and it will not be wrong to say that air transport cannot survive without the use of artificial intelligence. There are numerous functions in the machines and management processes that are controlled by the Al. There are a lot of features from booking the tickets to the takeoff and operation of the flights that AI takes care of. Al applications make air transport efficient, fast, safe, and provides a comfortable journey to the passengers, Changed the face of gaming: These days, we are able to play TV and computer games on the whole new level; all thanks to the artificial intelligence application. We all remember the time when “Super Mario” was the only game considered as the best. Now there are various gaming bots introduced and these will play with you and provide you the great entertainment. Virtual reality is also one example o artificial intelligence. A great future ahead as well: Engineers are designing machines by studying the human brains and trying to replicate the human intelligence. Although, it is one of the great challenges in engineering, but still a lot has been accomplished from where we began the. journey of artificial intelligence, There are numerous benefits. of understanding and replicating the brand in artificial intelligence. Apart from treating the brain injuries, diseases, advancements in communication technology, computer simulations, understanding the brain will help to design machines that have a more powerful impact towards the society. Even today we see machines that are capable of performing voice recognition, provide the response to human prompts, monitor and sense the hunian activities on a day to day basis. However, the future holds much more than what we are able to accomplish today. A dream to reinvent the world: Scientists are "ding on the back of AI when machine intelligence will surpass the human intelligence. Scientists believe that once the AI system starts working in its full capacity, it will reinvent the world that we know today. Think of the world where all the menial tasks Such as garbage disposal, construction, digging and so on will be taken care of by the AI application. It will be a time when the hierarchical order dictates the limits of a human, 'twill be the world where no one will be looked down upon and every human will be sinsidered equal. In this way, the humans can then focus their strengths on higher levels ofwork to accomplish a lot more and always taking technology to the new heights. Artifical Intelligence plays a very important role in not just the development: ines ‘td processes but also the humans to the next level. With the rapid growth in technol logy ‘evelopment, we can expect a lot more exciting features and uses of AI in the a 22.(b) Discuss issues in design of search program. 7 ( 7 Ans, Search is one of the operational tasks that characterize AI programs best. Best every AI program depends on a search procedure to perform its preseribed {ions Probloime eve typically dofmed in terms of states, and solutions correspond a —UmUr to searching through the different states Uli one = Solving a problem then ae st more are found. ii cayete set of gt ne ieie pes ible via some sequence of actions the agent takes is fete acne crs Which ate possible via Ba t the agent actually performs is “Sbeargy, Careh space. The series of actions that ined property. s ion i the required prop’ i Path, and the final state is a solution if it has ‘ay tha initial atate, the “Wetonwtl Considerations for Search: If we can Coe oe to start haw to “nd the goal check for a search problem, then we Scanned with CamScannerSixth Semester, Artificial Intelligence 6-2018 i ‘This leaves the important question of how to choo move and when to stop i> Se . ae ee daeiceca Wel ae which operator (oP strategy. Before we worry about © the fallowing need to be taken into consideration + Path or Artefact Broadly speaking, there are two different reasons: to undertake a se artefact (a particular state), or to find a path from one given * Completeness tly what strategy to use, areh: to find an, ate to another given state, It’s also worth trying to estimate the number of solutions to a problem, and the density of those solutions amongst the non-solution Ifour search strate; that it is complete, + ‘Time and Space Tradeoffs is guaranteed to find all the solutions eventually, then we say In practice, you are going to have to stop your agent at some stage if it has not found a solution by then. Hence, if we can choose the fastest search strategy, then this will explore more of the search space and increase the likelihood of finding a solution, Al practitioners often have to devise clever ways to trade memory and time in order to achieve an effective balance. * Soundness Soundness in theorem proving means that the search to find a proof will not succeed if you give it a false theorem to prove. This extends to searching in general, whore search is unsound if it finds a solution to a problem with no solution, This kind of unsound search may not be the end of the world if you are oul interested in using it for problems where you know there is a solution (and it performs well in finding such solutions). Another kind of unsound search is when a search finds the wrong solution to problem. This is more worrying and the problem will probably lie with the goal testing mechanism. * Additional Knowledge in Search. ‘The amount of extra knowledge available to. your uninformed search strategies, where no additional searches, where any information about the goal, it can be used to improve the efficiency of the search strategy. Q. 8. (a) Explain the various categories of Production System, Ans. A production system (or production rule s: typically used to provide some for Syste agent will affect how it performs. knowledge is given, and heuristic intermediate states and operators @) stem) is a computer program! 'm of artificial intelligence, Components of Produstiot Production rules Knowledge base (KB) Control Strategy- Order in whi i i Knowledge base rate ‘#y- Order in which rule is applied compared to Database! vale Renan checks current state with LHS ofrules in KB and finds appropriat® Four classes of production systems: 1. Amonotonic production system 2. Anon- monotonic production system 8. A partially commutative production system 4.A.commutative production system, Scanned with CamScannerLP. University {BCA} Akash Books Categories of Production System Nommonotonic. [Robot navigation Brege Production systems describe the operations that can be performed in a ee Monotonic production system:-A system in which the application of a rule never prevents the later application of another rule that could have siso been applied at the time the first rule was selected ,r—~—r—r——..__si“‘i ‘SUNN the application of a particular sequence of ru transforms state X into state Y, then any Permutation of those rules that is allowable also transforms state x into state ¥. Theorem Proving falls under monotonic partially communicativ. system. Blocks world and 8 puzzle problems like chemical analysis and synthesis come under monotonic, not partially commutati ms. Playing the game of partially commutative system. For any problem, several production systems exist. Some wi others. Though it may seem that there is no relationship between kinds of problems and kinds of production systems, in practice there is a definite relationship. Partially commutative, monotonic production systems are useful for solving ignorable problems. These systems are important for man'implementation standpoint because they can be implemented without the ability to backtrack to frevious states, ‘hen it is discovered that an incorrect path was followed. Such eystems increase the ‘isiency since itis not necessary to keep track ofthe changes made in the search process, Monotonic partially commutative sy ystems are useful for problems in which anges occur but ean be reversed and in which the order of operation is not critical (ex: 3 puzzle problem). idge comes under non monotonic, not Production systems that are not partially commutative are useful for: many problems ‘shich irreversible changes occur, such as chemical analysis. When dealing with such 222s, the order in which operations are performed is very important and hence correct. ns have to be made at the first time itself. 7 3. (b) What are the elements of AI? Diseuss various Application area a Ans. The basi st important elements of Artificial Intelligenee(AI). However, Fanos eeeseeer aa sub fields of Al depends on the contest. e.g.:-can Lan: us, Machine Representation & Learning (ML), Reinforcement Learning, Deep ag ete. 1 Problem Solving mostly deals with constraint satisfaction problems, Solve by Starch 2 Conzin knowledge representation, reasoning of it and plan of execution. First . = logicis dealt here. ing. This deals with “iain knowledge representation and its reasoning. This ‘ilistic graphical models and decision making» ility distributions wan tng, ML algorithms, Neural Networks, probability distrib Scanned with CamScannerSixth Semester, Artificial Intelligence ae ‘This involves parsi ae jon taken by agents. This involves parsing of 5, Perception, communication and action 80) ning fora robotic agent, * texts, Representation of images (can be in 3), i jitectures. a ‘Applications of Artificial Intelligence _————_ ° Antificial Inteligence | it Natural ence. lore Interface ; Applications Applications Applications © Expert Systems © Visual Perception ® Natural Languages © Learning Systems © Taclilily © Speech Recognition © Fuzzy Logic © Dexterity © Multisensory Interfaces Genetic Algorithms * Locomotion ® Virtual Reality © Neural Networks ° Navigation * Intelligent Agents UNIT-IL Q.4. (a) What are the qualities of a good Knowledge Representation System? ) Ans. The following properties should be possessed by a knowledge representation system. Representational Adequacy —The ability to represent the required knowledge; Inferential Adequacy - The ability to manipulate the knowledge represented to produce new knowledge corresponding to that inferred from the original; Inferential Efficiency - The ability to direct the inferential mechanisms into the most productive directions by storing appropriate guides; Acquisitional Efficiency - The ability to acquire new knowledge using automatic methods wherever possible rather than reliance on human intervention. There are several types of schemas that have proved useful in Al programs. They include (i) Frames: - Used to describe a collection of attributes that a given object possesses (e.g: description of a chair). GiSeripts: - Used to describe common sequence of events (e (iii) Stereotype: (v)Rule models: a production system, 2: 4,0) What is predicate logic? How knowledge Representation can be (6.5) 4 restaurant scene): ~ Uned to describe characteristics of people, Used to describe common features shared among a get of rules i? achieved using predicate logic? Mustrate, Ans. Predicate logic is an extensi have everything that exists in propositi ion of propositional logic. In predicate logic, ¥" ional logic, but now you have the ability to attrib! Scanned with CamScannerLP. University-(BCA]-Akash Books »perties and relationshi dnd quantification Al-place predicate is a statem: An example-1 of this would be “Two is an even number.” We can also use variables that range over objects, but aren’t names of specific things themselves. An example of that would be “xis an even number.” ent that says something about an object, * P?x=x is an even number * Let a= two, then P! a= two isan even number “These 1-place predicates can be connected with the connectives from propositional Now example-2, let’s sa Everything is blue, To indicate the idea of e We usually say, “There exists a thing that is blue.” ‘The symbol we use for there exist is 3, 'y you wanted to say that ‘verything we use the symbol V. ‘You can also relate one object with another (or itself) using a multi-place predicate, Like “x respects y,” or “x is between y and z.” The difference ig the superscript used and how many variables you place on the right side of the capital letter (2- Predicate R°xy or a 3-place predicate Bxyz), Prediéate logic: Example * P(x)-> Pis predicate of x. * Teacher (sks) -> sks is a teacher (in loosely term) ~>teacher is predicate of sks. * All indoor games are easy = Vx:indoor_game(x) -> easy(x) * Rajiv likes only cricket Like (Rajiv, Cricket) * Everyone is loyal to someone Ny: loyaltota.y) ' rome Differentiate between Inheritable Knowledge & a : ge, Ans, 5 Repre Atta ‘ferential Knowledge Sent knowledge as formal logic: "8s have tails yx « dog(s) + hasatail(x) Advantages: Scanned with CamScanner10-2018 Sixth Semester, Artificial Intelligence * Asset of strict rules. © Canbe used to derive more facts. © Truths of new statements can be verified. d correctness. : : an ete procedures available to in implement standard rules of logic. + Popular in Al systems. e.g Automated theorem proving. Inheritable knowledge Relational knowledge is made up of objects consisting of * attributes * Corresponding associated values. We extend the base more by allowing inference mechanisms: * Property inheritance o Elements inherit values from being members of a class. © Data must be organised into a hierarchy of classes (Fig). age Adult Male |—sf 35 isa Musician Isa cae ‘Avant Gardel Jazz instance instance ae nds: bands Miles Davis Group Naked Ci i Miles Davis Quintet . Waessede Fig. Pi 0 presenting knowledge (f wath two kinds ofentitien eu Mall variety of know ge (facts) A. Pacts: Truths in some re 8 elevant world. Thes B. Representatio: . ns of facts in e n actually be able to manipulate. formalism, ‘Phe One wa 10 think of structuring hese ey level, at which oh facts are described, and Ob) the tities is at two | levels: (wu) th edge se ayMol Hove a ene C4) the know Which representations of Scanned with CamScannerAY, University IBCALAkash Books 2018-11 atgeta at the Knowledge loved ane a jy pngrams. ‘The fiets and representations are linked with two-way mappings. This linkis ealled presentation mappings. ‘The foward representation mapping maps from facts to representations, ‘Tho backward represontation mapping goes the ather w representations to faets, lofined in terms of symbols that can be manipulated from One common representation is natural Lang Regantloss af he representation tar tiets wou emoornod with am English repre: information into and ont of the sentences to the rey .esent se (particularly English) sentences, ina program, we may also need to be ‘ntation of thoso facts in order to facilitate getting system, We need mapping functions from English ation we actually use and from it back to sentences, UNITAIT sanguage Processing. Explain various types of ) is (NLP) is a subfield of Artificial Intelligence that is Heused on enabling computers to understand and process human languages, to get Computers closer ton human-level understanding of language. Computers don’t yet have the same intuitive udorstanding of natural language that humans do, ‘They can't really understand what the language is really trying to Recent advances in Machine Learning (ML) have enabled computers to do quit8.a {ot of useful things with natural language! Deep Learning has enabled us ea write params fo perform things like Iingnage translation, semantic understanding, and text summarization, All of these things add real-world value, making it easy for you to understand and perform computations on large blocks of text without the manual effort, The component forms of knowledge needed foran understanding of natural language are sometimes classified according to the following levels, * Phonological - Thisis knowledge which relates sounds to the words we recognize, + Morphological- This is lexical Knowledge which relates to word constructions, Syntactic- This knowledge relates to how words are put together or structured to form grammatically correet sentence in the language. Semantic- ‘This knowledge is concerned with the meanings of words and phrases and how they combine to form sentence meanings. Pragmatic- This is high-level knowledge which relates to the use of sentences in different contexts and how the contest affects the meaning of the sentences. World- World knowledge relates to the language a user must have in order to understand and carry on a conversation. It must include an understanding of the other person’s beliefs and goals. ; The following are common types of natural language Processing. 1. Optical Character Recognition- Converting written or printed text into-data. 2 Speech Recognition: Converting spoken words into data. + Machine Translation- Translating text from one language to te Fer, Natural Language Generation- Formatting es erase co ‘ample, a Weather system that generates weather ste ee an ‘Sietgin tient Analysis- Interpreting basic Serena care Siniek brand Fat metres ty whether comments are postive or Aer a brand and how Nog, Of thay netics might look at how many people are Sttalk is positive, Q.6, (a) Deseribo Natural NLP Techniques. Ans, Notunal Language Pry: Scanned with CamScannerWWE Ninth Nomenton, Artificial Intelligence Semantic Search: Underatanding and anawering questions posed in a natural 6, Semantic Searvhy Unde Tangnage. Ror example of that bar T went to Laat weekend.” 7. Machine Leaminys can be learning related t 8. Natural Language customize computer pry an app that snyggests: new experiences who have sieves career.” Vaing uatueal language to train artificial intelligence, This Jangnaye itaelfor another topic such as economics, fo Viogramming: Tools that allow end users to create of amis with natural language, Por example “make mg carvers paths to me based on people with similar skills ang fully improved thoir salary and quality of life with a new 9. Affective Computing: is the machine equivalent of emotional intelligence. Allows user interfaces to understa like conversations, vwately procomaingg a search query such as “give me the address | nd human ideas and emotional states and to take on human. Q.6.(b) What is learning? Explain how learning is helpful for AI? (6.5) Ans, Machine loarning isan application ofartifi system oxpl programs that can access The process ofl experience, or Instruction, i inthe future based an the e: computers learn automati actions accordingly, ning begins with observations or data, s ail intelligence (AT) that provides tho ability to automatically learn and improve from experience without being tly programmed, Machine learning focus son the development of computer data and use it learn for themselves, uch as examples, direct i onder to look for patterns in data and make better decisions samples that we provide. The primary aim is to allow the ly without human intervention or assistance and adjust Some machine learning methods are helpful for Al Machine learning algorithms are often eat targets for any new input compare its output with th the model accordingly. * In contrast, information used to train is neit how systems can infer a fu The system doesn't figure inferences, from datasets t Semi- ised learning, since they and relevant resources in ord: * Reinforcement m: interacts with its environ: Trial and error search ani reinforcement learning. Supervised machine learning algorithms can in the past to new data using labeled example: the analysis of a known training dataset, the learni function to make predictions about the output values. The ment by producing egorized as supervised or unsupervised. apply what has been learned s to predict future events, Starting from ng algorithm produces an inferred system is able to provide nt training. ‘The learning algorithm can also ‘correct, intended output and find errors in order to modify after sufficie: unction to describe a hidd. out the right output, but it explores the data and can draw 'o describe hidden structures from unlabeled data. machine learning algorithms fall Somewhere in between use both labeled and unlabeled data erally doesn't require additi achine learning algorithms is a learning method that actic “ 7 : id delayed reward are the and discovags errors or rewards. This method allow Scanned with CamScannerLP, University INCAL-Akauh Books OWA Hy determine the ideal bohaviour within a wpocifi ximize its performance, Simple reward foodback in required for the Hgent to learn ch action is best; this is known as the vointorcoment signal Machine learning enables analysis of masaive quauntition of data, While it pon dolivers tistor, more accurate results in context In order to ally rdor to identify profitable apportunition or dangerous risks, it may also require nddit MI Limo and rosoureen to tain It proporly, Combinitig machine learning with Aland usnitivo (ochnologion enn mul iLeven more effective in processing largo volumen of inlie mation, Q.7. (a) Comparison botween Syntactic Procossing & Somante Pre (6.5) ‘This knowledge relates to how words ure pul Cogothor or structured mmatically [ Lexical Analysis & Syntactic Analysis | [ Semantic Analysis ) caries Veccucren f Disclosure Integration * Syntactic Analysis (Parsing): It involves analysis of words in the sentence for grammar and arranging words ina manner that shows the relationship among the words, The sentence such as “The school gocs to boy” is rejected by English syntactic analyzer, i aws the ex: ing or the dictionary meaning from *_Semantic Analysis : It draws the exact meaning ort n m the text. The text is checked for meaningfulness. It is done by mapping syntactic Structures and objects in the task domain. The semantic analyzer disregards sentence Such as “hot ice-cream”, : Implementation Aspects of Syntactic Analysis | ' There are a number of algorithms researchers have developed for syntactic analysis, “econsider only the following simple methods — * Context-Free Grammar * Top-Down Parser Contet-Fy . eee ees ft-hand side of i ith asi mbol on the le the tthe grammar that consists rules with a single ayrbol Teurite Tules, Let us create grammar to parse a sent bird pecks the grains* "eles (DET) saan | the but Scanned with CamScanner14-2018 Sixth Semester, Artificial Intelligence Nouns : bird | birds | grain | grains Noun Phrase (NP) : Article + Noun | Article =DETN | DETADJN Verbs : pecks | pecking ! pecked Verb Phrase (VP): NP V | VNP ‘Adjectives (ADJ) : beautiful | small | chirping ‘The parse tree breaks down the sentence into structured parts so that the computer can easily understand and process it. In order for the parsing algorithm to construct this parse tree, a set of rewrite rules, which describe what tree structures are legal, need to be constructed. . ‘These rules say that a certain symbol may be expanded in the tree by a sequence of other symbols. According to first order logic rule, if there are two strings Noun Phrase (NP) and Verb Phrase (VP), then the string combined by NP followed by VP is a sentence. ‘The rewrite rules for the sentence are as follows — SNP VP NP» DETN | DETADJN VP > VNP Lexocon ~ DET +a | the ADJ —> beautiful | perching N > bird | birds | grain | grains V-— peck | pecks | pecking ‘The parse tree can be created as shown — 4+ Adjective + Noun Demerits — They are not highly . ! ly precise. syntactically correct according to parser pene as a correct sentence. ‘Scanned with CamScannerLP. University-{[BCA}Akash Books 2018-15 ~ To bring out high precision, multiple sets of grammar need to be prepared. It may require a completely different sets of rules for parsing singular and plural variations, passive sentences, ete., which can lead to creation of huge set of rules that arc unmanageable. Top-Down Parser Here, the parser starts with the S symbol and attempts to rewrite it into a sequence of terminal symbols that matches the classes of the words in the input sentence until it consists entirely of terminal symbols. ‘These are then checked with the input sentence to see if it matched. Ifnot, the process is started over again with a different set of rules. This is repeated until a specific rule is found which describes the structure of the sentence. Merit - It is simple to implement. Demerits — ~ Itis inefficient, as the search Process has to be repeated if an error occurs. > Slow speed of working, Semantic Processing: When. you really want to understand meaning in text What's really difficult is understanding what is being said, and doing it at scale. For humans, the Way we understand what's being said is almost an ‘unconscious Process. To understand what a text is talking about, we rely on what we already know about language itself and about the concepts present in a text. Machines can't rely on these same techniques. "Some technologies only make you think they understand text. An approach based on keywords or statistics, or even pure machine learning, may be using a matching or frequency technique for clues as to what a text is “about.” These methods ean only go so far because they are not looking at meaning. Semantic analysis describes the process of understanding natural language the way that humans communicate-based on meaning and context. Look at how a cognitive technology like Cogito performs semantic analysis, The semantic analysis of natural language content starts by reading all of the words in content to capture the real meaning of any text. It identifies the text slements and assigns them to their logical and grammatical role, Itanalyzes context in the surrounding text and it analyzes the text structure to accurately disambiguate the Proper meaning of words that have more than one definition. Semantic technology processes the logical structure of sentences to identify the most televant elements in text and understand the topic discussed. It also preci the “letionships between different concepts in the text, For example, Paneer that plstis about “polities” and “economies” even ifit doesn't contain ieeeeaere 'ttrelated concepts such as “election,” “Democrat,” “speaker of the house, get, * or “inflation.” machines aye otse semantic analysis and natural Tanguage De ce caisitag ing’™tically understand text, See ou ae aera aoe ine Coe ering tion-that potentially valuable piece Seis ieee for customer support, Sustomer service log-into the realm of busin ’ Porate intelligence or knowledge management. Scanned with CamScanner16-2018 Sixth Semester, Artificial Intelligence Qa Ans, Discourse Integr i just meaning of the seritence j i iately succeeding sentence. : re ‘meaning of Individual sentence is depending on previous sentence. Ex: a tic processing. @) in discourse & pragmat (b) Explain gration : The meaning of any sentence depends upon the before it. In addition, it also brings about the meaning of Bill had a red balloon. John wanted it. ~ Pragmatic Analy: uring this, what was said is re-interpreted on what it actually meant. It involves deriving those aspects of language which require real world knowledge. Understanding the text & dialogues. It derives Knowledge from external common sense information. Ex: Do you know what time it is? It does not mean the speaker asking you the time!!! We should understand what to do? UNIT-IV Q8. Why LISP is considered to be appropriate language for AI technique? Write a LISP program to print factorial of a given number. (12.5) Ans, LISP in facilitat language. For instance, LISP has a unique ton heh ay oa aandardAl "m which facilitates exploration ntelligence. is more efficient in solving specific is writing. It is highly suitable in and implementation. of different lev lutions a develope: inductive logic projects and machine learning, Advantages Fast and efficient in coding as iti i | ah ‘ ng as itis supported by compilers instea | * LISP offers Specific control (write-line “Please enter an (setq x (read) (defun factorial(n) (if(=n1) over systems resultin, 'g to their maximum use. jumber...”) | | | | | | Scanned with CamScannerad LP. Univer jity-{ BCA]-Akash Books 2018-17 (setq a1) ) fond (wetq a n (factorial (1 1))) ) (format ) (factorial x) Q.9, What is Expert System? What are the characteristics of a good Expert system? Explain MYCIN expert system. (12.5) ‘Ans. An Expert System is a computer program that possesses or represents knowledge in a particular domain, has the capability of processing/ manipulating or reasoning, with this knowledge with a view to solving a problem, giving some achieving ortoachieve some specific goal. ‘An expert system may or may not provide the complete expertise or functionality of human expert but it must be able to assist a human expert in fast decision making. ‘The program might interact with a human expert or with a customer directly. “-D! is ~D" na) ‘The basic properties of an expert system: © Ittries to simulate human reasoning capability ubout a specific domain rather than the domain itself, This feature separates expert systems from some other familiar programs that use mathematical modelling or computer animation. In tun expert system the focus is to emulate an expert’s knowledge and problem solving capabilities and if possible, at a faster rate than a human expert, + Itperforms reasoning over the acquired knowledge, rather than merely performing some calculations or performing data ret eval, ny heuristic or approximate models which, unlike 1r¢ not guaranteed to succeed. 1YCIN, one of the earliest designed expert systems © Itean solve problems by u: other algorithmic solution: MYCIN (An expert system): in Stanford University in 1970s. MYCIN’s job was to diagnose and recommend treatment for certain blood infections. Wo the proper diagnosis, itis required to grow cultures of the infecting organism which isa very time consuming process and sometime patient is in a critical sti ite. So, doctors i have toeeme up with quick guesses about likely problems from the available data, and ' we these guesses to provide a treatment where drugs are given which should deal with ‘ny type of problem. So MYCIN was developed in ord er to explore how human experts make these rough important) guesses based on partial information. Sometimes the problem takes *ccther shape that an expert doctor may not available every-time every-where, in that “uation also an like MYCIN would be handy. \d expert system like : ; dagY GIN represented its knowledge as & set of IF-THEN rules with certainty factors. tthe MYCIN rule could be like: 4 cl lection is primany-bacteremia AND the site of ‘the culture is one of the sterile intestinal tract. roid infection occurred. - Bs _O the suspected portal of entry is the gastroi | HEN there is suggestive evidence (0.8) that bacte Scanned with CamScanner18-2018 Sixth Semester, Artificial Intelligence The 0.8 is the certainty that the conclusion will be true given the evidence, try evidence is uncertain the certainties of the pieces of evidence will be combined with certainty of the rule to give the certainty of the conclusion. Ne MYCIN has been written in Lisp, and its rules are formally represented ag lig expressions. The action part of the rule could just be a conclusion about the probler being solved, or it could be another lisp expression. 2 MYCIN is mainly a goal-directed system, using the backward chainin reasoning approach. However, to increase it reasoning power and efficiency MYCIN alse uses various heuristics to control the search for a solution. One of the strategies used by MYCIN is to first ask the user a number of predefined questions that are most common and which allow the system to rule cut totally unlikely diagnoses. Once these questions have been asked, the system can then focus on particular and more specific possible blood disorders. It then uses backward chaining approach to try and prove each one. This strategy avoids a lot of unnecessary search, and is similar to the way a doctor tries to diagnose a patient. The other strategies are related to the sequence in which rules are invoked, One of the strategy is simple i.e., given a possible rule to use, MYCIN first checks all the antecedents of the rule to see if any are known to be false. If yes, then there is no point using the rule. ‘The other strategies are mainly related to the certainty factors. MYCIN first find the rules that have greater degree of certainty of conclusions, and abandons its search _ once the certainties involved get below a minimum threshold, say, 0.2. ‘There are three main stages fo the interaction with MYCIN. In the first stage, i | datz ebout the cate is gathered go the system can come up with a broad diagnosis. In the second more directed questions are asked to test specific hypotheses. At the end of this section it proposes a diagnosis. In the third stage it asks questions to determine an appropriate treatment, on the basis of the diagnosis and facts related to the patient. After that it recommends some treatment. At any stage the user can ask why a question was asked or how a conclusion was reached, and if a particular treatment is recommended the user can ask if alternative treatments are possible. | Scanned with CamScanner