APPLICATIONS Al applications ~ Language Models — Information Retrieval- Information Extraction ~ Natural Language Processing ~ Machine Translation ~ Speech Recognition - Robot ~ Hardware Perception ~ Planning ~ Moving\ Al or artificial intelligence is the simulation of human intelligence processes by machines, especially computer systems. These processes include + Leaming + Reasoning + Self-correction Some of the applications of Al include expert systems, speech recognition and machine vision. Artificial Intelligence is advancing dramatically. 5.1 APPLICATIONS OF ARTIFICIAL INTELLIGENCE 1) Artificial Intelligence in Healthcare: Companies are applying machine learning to make better and faster diagnoses than humans. One of the best-known technologies is IBM’s Watson. It understands natural Janguage and can respond to questions asked of it. The system mines patient data and other available data sources to form a hypothesis, which it then presents with a confidence scoring schema. Al is a study realized to emulate human intelligence into computer technology that could assist both, the doctor and the patients in the following ways: ° By providing a laboratory for the examination, representation and cataloguing medical information ® By devising novel tool to support decision making and research ¢ By integrating activities in medical, software and cognitive sciences * By offering a content rich discipline for the future scientific medical communities. 2) Artificial Intelligence in business: Robotic process automation is being applied to highly repetitive tasks normally performed by humans. Machine leaning algorithms are being integrated into analytics and CRM (Customer relationship§2 _Aniificiat intelligence — management) platforms to uncover information on how to pee ays custome Chat bots have already been incorporated into websites and e-companies tg provide immediate service to customers. Automation of job positions has also become a talking point among academics and IT consultancies. 3) AI im education: [t automates grading, giving educators more time. It can also” assess students and adapt to their needs, helping them work at their own pace. 4) Al in Autonomous vehicles: Just like humans, self-driving cars need to have sensors to understand the world around them and a brain to collect, Processes and choose specific actions based on information gathered. Autonomous vehicles are with advanced tool to gather information, including long range radar, cameras, and LIDAR. Each of the technologies are used in different capacities and each collects different information. This information is useless, unless it is processed and some form of information is taken based on the gathered information. AI has several applications for these vehicles and among them the more immediate ones are as follows: Directing the car to gas station or recharge station when it is running low on fuel. Adjust the trips directions based on known traffic conditions to find the quickest route Incorporate speech recognition for advanced communication with passengers. Natural language interfaces and virtual assistance technologies. 5) Al for robotics will allow us to address the challenges in taking care of an aging population and allow much longer independence. It will drastically reduce, may be even bring down traffic accidents and deaths, as well as enable disaster response for dangerous situations for example the nuclear meltdown at the fukushima power plant ©) Cyborg Technology: One of the main limitations of being human is simply our own bodies and brains. Researcher Shimon Whites on thinks that in the future, we will be able to augment ourselves with computers and enhance many of our own natural abilities. Though many of these possible cyborg enhancements would be _ added for convenience, others may serve a more practical purpose. Yoky Matsuka of Nest believes that AI will become useful for People with amputated limbs, as the brain will be able to communicate with a robotic limb to give the patient more | control. This kind of cyborg technology would significantly reduce the limitations | that amputees deal with daily. : 5.2 LANGUAGE MODELS Formal grammars («.g. regular, context free) give a hard “binary” model of the legal sentences in a language. NLP Probabilistic model of a language that gives a probability string is 2 member of @ language or not. To specify a comect probability distribution, tol.Applications 53 aoe . Uses of Language Models + Speech Recognition + OCR & Handwriting Recognition * Machine Translation * Generation + Context sensitive spelling correction. A language model also supports predicting the completion of a sentence Predictive text input systems can guess what is been typed and provide choices on how to complete it N- Gram Word Models + This model is considered over sequences of words, characters, syllables or other units. + Estimate probability of each word given prior context. + An N-gram model uses only N-1 words of prior context. Unigram: P(phone) Bigram: P(phone | cell) ‘Trigram: P(phone | your cell) + The Markov assumption is the presumption that the future behavior of a dynamical system only depends on its recent history. In particular, in a Kth-Order Markov Model, next state only depends on the k most recent states, therefore an N — gram model is a (N-1) — order Markov model. N-gram Character Models + One of the simplest language models: P(c:') + Language identification: given the text determine which language it is written in. * Build a trigram character model of each candidate language: P(¢, !¢,-2,.1-!) + Train and Test Corpora * A language model must be trained on a large corpus of text to estimate good parameter values. * Model can be evaluated based on its ability to predict a high probability for a disjoint test corpus. ; ° The training corpus should be representative of the actual application data. : * To handle words in the test corpus that did not occur in the training data an explicit symbol is used. ‘Symbol to represent unknown words () Preplexity ~ Measure of how well a model “fits” the test data.§4_ Artificial Intelligence PerplexithW,® ) oie as Pow 7 a0y) Smoothing - reassigns probability mass to unseen events; 3. INFORMATION RETRIVAL What is Information Retrieval? Information retrieval (IR) is finding material (usually documents) of an unstructured nature (usually text) that satisfies an information need from within larg collections (usually stored on computers).Generically, “collections”,Les: used, “corpora” are searched and “documents” namely web pages 's, PowerPoint slides, paragraphs, ete. are retrieved. Information Retrieval system consists of Software program that facilitates a user _in finding the information the user needs, The Information Retrieval System was coined by Calvin Mooers in 1952. These information retrieval systems were, truly speaking, document retrieval system, since they were designed to retrieve information, Information retrieval deals with storage, nization and access to text, as well as multimedia information resources. Information Retrieval is a process of searching some collection of documents, using the term document in its widest sense, in order to identify those documents which deal with a particular subject. Any system that is designed to facilitate this literature searching may legitimately be called an information retrieval system. Information retrieval systems originally meant text retrieval system dealing with textual documents, modem information retrieval multimedia information comprising text, audio, images and video. While many features of conventional text retrieval system are equally applicable to multimedia information retrieval, the specific nature of audio, image and video information have called for the development of many new tools and techniques for information retrieval Modern information retrieval deals with storage, organization and access to text, a8 well as multimedia information resources. since they were stems deal with The concept of information retrieval presuppos Q that there are some documents or records containing information that have been organized in an order suitable for easy retrieval. The documents or records we are concerned with contain bibliographic information which is quite different from other kinds of information or data. We may take a simple example. If we have a database of information pertaining to an office ot a supermarket, all we have are the different kinds of records and related facts, ike names of employees, their positions, salary, and 0 on, or in the case of a supermarket, names of different items, prices, quantity, and so on, The main objective of @ bibliographic information retrieval system, however, is to retrieve the information- either the actual information or the documents containing the information that fully oF Partially match the user's query, The database may contain abstracts or full texts of document, like newspaper articles, handbooks, dictionaries, encyclopedias, legal documents, statistiesete,, as well as audio, images, and video inforApplications _ 5.5 An information retrieval system thus has three major components- the document subsystem, the users subsystem, and the searching/retrieval subsystem. These divisions are quite broad and each one is designed to serve one or more functions, such as: Analysis of documents and organization of information (creation of a document database) Analysis of user’s queries, preparation of a strategy to search the database + Actual searching or matching of users queries with the database, and finally + Retrieval of items that fully or partially match the search statement. An IR isa 3 step Process: 1) Asking a question (how to use the language to get what we want?) 2) Building an answer from known data.(How to refer to a given text?) 3) Assessing the answer. (Does it contain the information we are seeking.) Information Retrieval Cycle Source seecen ros Selection — Documents Systemaiscovery ‘ Vocabulary aiscovery = Loe: Delivery Document collection indexer__ Applications 5) artel + The basic unit of data is a document (can be a file, 4 paragraph, ete.) + A document corresponds to free text (may be unstructured) + Allthe documents are gathered into a collection (or corpus) + Example: 1 million documents, each counting about 1000 words if cach word is encoded using 6 bytes 109 x 1000 * 6/1024 = 6GB Components of Information Retrieval In an information retrieval system there are the documents or sources of information on ‘one side and on the other there are the user’s queries, These two sides are linked through a series ‘of tasks. Lancaster mentions that an information retrieval system comprises six major subsystems: * The document sul + The indexing subsystem + The vocabulary subsystem ‘+ The searching subsystem + The service-system interface,and + The matching subsystem Information r Organized sources {> | Represemation | > | tntormation The broad outline of an information retrieval system apa dc (Eaeemem ! Information Matching Users Ey [ Query Analysis ==) Analysed Queries Three major components of IRS 1) Document subsystem a) Acquisition b) Representation ) File organization 2) User sub system: a) Problem§8_Anificial Intelligence b) Representation. ©) Query 3) Searching /Retrieval subsystem a) Matching b) Retrieved objects Acquisition (Document subsystem) + Selection of documents & other objects from various web resources, + Mostly text based documents full texts, titles, abstracts © but also other objects: © data, statistics, images, maps, trade marks, sounds ... * The data are collected by web crawler and stored in data base Representation of documents, objects(document subsystem) + Indexing ~ many ways : © free text terms (even in full texts) © controlled vocabulary - thesaurus © manual& automatic techniques. + Abstracting; summarizing * Bibliographic description: ° author, title, sources, date... ° metadata + Classifying, clustering + Organizing in fields & limits © Basic Index, Additional Index. Limits File organization (Document subsystem) + Sequential © record (document) by record + Inverted ° term by term; list of records under each term + Combination + indexes inverted, documents sequential + When citation retrieved only, need for document filesApplications 5.9 + Large file approaches + for efficient retrieval by computers Problem (user subsystem) + Related to users’ task, situation vary in specificity, clarity « Produces information need © ultimate criterion for effectiveness of retrieval i,chow well was the need met? «Information need for the same problem may change, evolve, shift during the IR process adjustment in searching «often more than one search for same problem over time Representation ( user subsystem) + Converting a concept to query + What we search for. + These are stemmed and corrected using dictionary + Focus toward a good result + Subject to feedback changes Query - search statement (user & system) + Translation into systems requirements & limits © start of human-computer interaction © query is the thing that goes into the computer + Selection of files, resources + Search strategy - selection of: © search terms & logic « possible fields, delimiters © controlled & uncontrolled vocabulary » variations in effectiveness tactics + Reiterations from feedback © several feedback types: relevance feedback, magnitude feedback.. query expansion & modification Matching - searching (Searching subsystem) + Process of matching, comparing{10 Anificlal tmteltigence —— he * search: what documents in the file match the query ? * Various seareh alporithms: © exaet mateh - Boolean * still available in most, if not all systems: best match » ranking by relevance * increasingly used ep. on the web hybrids incorporating both = og. Target, Rank in DIALOG + Bach has strengths, weaknesses No ‘perfect’ method exists and probably never will Retrieved documents -from system to user (IR Subsystem) * Various order of output Last In First Out (LIFO); sorted ranked by relevance ranked by other characteristics + Various forms of output When citations only: possible links to document delivery Base for relevance, utility evaluation by users * Relevance feedback Kinds of information retrieval systems Two broad cat fortes of information retrieval system can be identified in- house and online In- House information retrieval systems are set up by a particular library or information center to serve mainly the users within the organization, One Particular type of in-house database is the library catalogue. Online public access catalogues (OPACs) Provide facilities for library users 10 cary out online catalogue searches, and then to check the availability of the item required. Online IR is nothing but retrieving data from web sites, web pages and servers that may include data bases, images, text, tables, and other types, Functions of information retrieval system An information retrieval system deals with variou Ws sources of information on the one hand and user's requirements on the other. It mustApplications _ 5.1 + Analyze the contents of the sources of information as well as the user's queries, and then + Match these to retrieve those items that are relevant The major functions of an information retrieval system can be listed as follows: + To identify the information (sources) relevant to the areas of interest of the target users community * To analyze the contents of the sources (documents) + To represent the contents of the analyzed sources in a way that will be suitable for matching user's queries + To analyze user's queries and to represent them in a form that will be suitable for matching with the database + To match the search statement with the stored database + To retrieve the information that is relevant, and + To make necessary adjustments in the system based on feedback form the users. Features of an information ret Liston and Schoene suggest that an effective information retrieval system must have provisions for: val system + Prompt dissemination of information + Filtering of information + The right amount of information at the right time + Active switching of information + Receiving information in an economical way + Browsing + Getting information in an economical way * Current literature + Access to other information systems + Interpersonal communications, and + Personalized help. Indexing usually consists of the several phases + After word segmentation, stopwords are removed. * These common words like articles or prepositions contain little meaning by themselves and are ignored in the document representation. * Second, word forms are transformed into their basic form, the stem. + During the stemming phase, ¢.g. houses would be transformed into house.{12 _Anificiat imsetiigence * For the document representation, different word forms are usually not necessary * The importance of a word for a document can be different. * Some words better describe the content of a document than others. ‘+ This weight is determined by the frequency off stem within the text of a document In multimedia retrieval, the context is essential for the selection of 8 form of query and Tepresentation. Different media representations may be matched against each other or transformations may become necessary (¢.g. to match terms against pictures or spoken language utterances against documents in writien text) + As information retrieval needs to deal with vague knowledge. exact proces methods are not appropriate. + Vague retrieval models like the probabilistic model are more suitable. + Within these models, terms are provided with weights corresponding to ther imponance for a document. + These weights mirror different levels of relevance. The result of current information retrieval systems are usually sorted bss of documents where the top results are more likely to be relevant according to the system. * In some approaches, the user can judge the documents retumed to him and tell the systems which ones are relevant for user. * The system then resorts the result set. * Documents which contain many of the words present in the relevant documents = ranked higher. + This relevance feedback process is known to greatly improve the performance + Relevance feedback is also an interesting application for machine learning . Based on a human decisions, the optimization step can be modeled with several approaches, e.g with rough sets. * In Web environments, a click is often interpreted as an implicit positive relevance judgment 54 INFORMATION EXTRACTION . Information extraction (IE) is the automated retrieval of specific information related © a selected topic from a body or bodies of text. + Information extraction tools make it possible to pull information from text documents. databases, websites or multiple sources. + IE may extract info from unstructured, semi-structured or structured, machine readable text. + JE is used imnatural language processing (NLP) 10 extract structured fom unstructured text,Applications _ 5.13 —_— s Information extraction depends on Named entity recognition (NER), a sub-tool used to find targeted information to extract. NER recognizes entities first as one of several categories such as location (LOC), persons (PER) or organizations (ORG) Once the information category is recognized, an information extraction utility extracts the named entity’s related information and constructs a machine-readable document from it, which algorithms can further process to extract meaning. IE finds meaning by way of other subtasks including co-reference resolution, relationship extraction, language and vocabulary analysis and sometimes audio extraction. Current efforts in multimedia document processing in IE include automatic annotation and content recognition and extraction from images and video could be seen as IE as well Because of the complexity of language, high-quality IE is a challenging task for artificial intelligence (Al) systems. What is IE? Main goals of IE + Flo pred Unstructured ae Structured Web Text Sequences oc wha did what fo whom and when? TE, Task & Subtasks + Releflonship extrac! Sate « Terminology extrocton feagurciemsml + Some concrele exampies eng frig. Plt tert sets fee onc orey ans + Coreference resaiution 5 Seortag mite Mew a oman ty Coeany LTE mys nag gre sain han area ee eTApplications _S.1S 56 MACHINE TRANSLATION Machine Translation (MT) is the task of automatically converting one natural language no another, preserving the meaning of the input text, and producing fluent text in the output Janguage. While machine translation is one of the oldest subfields of artificial intelligence research, the recent shift towards lar le empirical techniques has led to very significant improvements in translation quality Machine Transtation + €1) DAIYU ALONE ON BED TOR THINK pacenar sen lay there alone Dalyu's thoughts turned to Haochat + €3) CLUAR COLD PENETRATE CURTAIN | | ES coldness penetrated the curtains of her bed + ca NOT FEELING FALL DOWN TEARS Comte | | + G4: Almost without noticing It she had began to cry ‘ations of current MT computational models + Rough translation + Human post-editor - Computer-aided human translation (CAHT or CAT) © Software manual © PhD the: + Fully automatic, high-quality translation (FAHQT) mall sublanguage domains deals with Domains which are often found with limited vocabulary + Weather forecast © Cloudy with a chance of showers today © Outlook for Friday: Sunny + Air travel queries + Appointment scheduling + Restaurant recommendation + Recipes Challenges of MT ‘Typology: study of eross-linguistic similarities and differences + Morphology Syntax: order of verbs, subjects and objects6 Arifficial Intelligence Argument structure and linking, + Head-marking + Dependent-marking Verbs and satellite particles (direction, motion, Pronouns omission + Pro-drop English = [1] am reading a book, Spanish * Estoy leyendo un libro. + Referential density Cold: more inferential work to recover antecedents + Japanese, Chinese Hot: more explicit and easier + Spanish Other divergences: + Position of adjectives English: ,,green witch" Spanish: ,,witch green + Chinese relative clauses (in relation to English) | + Cultural aspects, ¢.g., calendars and dates » British English: DD/MM/YY * American English: MM/DD/YY ° Japanese: YYMMDD Lexical + Word sense disambiguation + Homonymy wall (Wand), wall (Mauer) + Polysemy «to know (knowing a fact) : wissen © to know (familiarity with a person/locatio kennen. + Grammar » Engli She likes to sing—— a Applications 5.17 ° German: ,,Sie singt gern.“ + Lexical gap © No ,privacy word in Japanese Classical MT + Direct — Word-by-word translation + Transfer — Source language parse structure to target language parse structure + Interlingual — Analysis using an abstract meaning representation Vauquois triangle Direct translation + Word-by-word translation + No intermediate structures, except for shallow morphological analysis — e.g., verb tenses, negations + Use of bilingual dictionaries + Followed by reordering rules ~ e.g., moving adjectives after nouns Transfer + Use of constractive knowledge — Differences between the two languages + Three phases ~ Analysis: parsing ~ Transfer + Lexical — English: home - German: nach Hause, Heim, Heimat, zu Hause, ete * Syntactic: - Nominal (ADJ NN) [ Nominal (NN ADJ) — Generation Direct + Transfer + Used by most commercial MT systems + Shallow analysis ~ Morphological analysis, part-of:speech tagging ~ Chunking (NPS, PPs, etc.) ~ Shallow dependency parsing + Transfer — Translation of idioms — Word sense disambiguation + Synthesis — Lexical translation (bilingual dictionary) — Reordering ~ Morphological generation18 Anificial Intelligence =e Interlinguat * Shortcoming of transfer ~ Transfer rules for each pair of rules ~ Not feasible op multilingual environments (e.g, EU) + Interlingual ~ Extract the meaning of the input and express it mie ee en Be Amount of knowledge transfer would be proportional to the number of language rather to the square ~ Deep semantic analysis from language X to the interlingual Fepresentation ~ Generating language Y from the interligual representation, Advantages: * Natural language processing components can be used for both the translation and generation Shortcomings: * Semantic analysis is hard and only feasible for sublanguage domains * Unnecessary disambiguation across many languages ©, Chinese has concepts for ELDER_ BROTHER and YOUNGER BROTHER Statistical MT * Model translation as the production of an output that maximizes Tepresent the importance of faithfullness and fluency Probabilistic models of faithfullness and fluency a value function that * Combine these models to choose the most Probable translation for a sentence §: best-translaion? “argmaxt fathfullness(T-, S ) fluency (1) Alignment in MT * Word alignment: ~ Mapping source words to target words in a set of paralle Fach target word comes from exactly one source word * Represent alignment as sequence of index numbers wAnd(1) the(2) program(3) has(4) been(5) implemented(6)* ,Le(1) Programme(2) a(3) éé4) mis(5) en(6) application(7)" A = 2,3,4,5,6,6,6 1 sentences 5.7 SPEECH RECOGNITION What is the definition of Speech Recognition? Speech Tecognition is the Process where a computer (or any other device) identifies spoken words. This is not only the recognition of natural language, but natural language in its verbal form. Speech recognition enables users to Control digital devices through voice commands instead of using conventional tools such as a I s board, buttons, mouse or keystrokes, Speech recognition allows you to provide input by simply talking.as : Applications _5.19 What are terminology of Speech Recognition? Artificial Intelligence (Al): The study and programming of computer systems that can perform tasks normally associated with human intelligence (including perception and discovery, decision-making, and speech recognition). Phoneme: The smallest unit of speech. Phonemes help distinguish one word from another, Algorithms: An equation or set of rules that a computer follows (coupled with Al decision-making capabilities) Artificial Neural Networks: These are considered the “foundation of Al,” and refer to the pieces of the processing system that computers use to perform human-like, intelligent problem- solving. Natural Language Processing (NLP): This is the area of Al that revolves around the study of machines interacting with humans through “natural” (i.e., human) language. Machine Learning: This is where artificial intelligence equips a machine to lea and improve its processes automatically without human intervention Interactive Voice Response (IVR): This is the technical name for the interactive telephone systems that can respond to voice and dial pad commands User Interface (UL): This is where the program “comes to the surface” where it can interact with a user. UI refers to all the means a system uses to interact with humans, including the desktop, display screens, keyboards, and mouse. Directed Dialogue: In contrast with a user speaking freely to the user interface, this is where specific, pre-determined phrases need to be utilized for the system to recognize the command and act on it. This is especially common in IVR systems on customer service telephone lines. Conversational Interface: The conversational interface is where the system intakes natural language, which in the case of speech recognition will always be spoken Speech Engine: This is the software that takes spoken input, compares it to available vocabulary and assigns meaning to it Deep Learning: This is a subset of machine learning, which is a subset of Al. It's considered “deep” because it has the ability to machine-learn from unstructured, unlabeled data on a partially or completely unsupervised basi What are the categories of Speech Recognition? Speech recognition can be categorized using the following paramete + Dependence on speaker + Recognition style~~ $20 Artificial Intelligence oo Dependence on speaker Speech recognition systems are classified depending on whether or not t they depend on the Speaker. This means that systems can be trained by and 0 a single person’s voi Ce, OF 10 a genera) Vocabulary regardless of the delivering voice. 1) Speaker-dependent 2) Speaker-independent Speaker-dependent * Speaker-dependent speech recognition systems require the unique biometric characteristics of a single person's voice in order to process the speech. ‘These systems are “trained” by the person who will ultimately be using them ‘New users must first “train” the software themselves by speaking to it so it can analyze the way they talk Speaker-dependent systems achieve higher accuracy of speech recognition than speaker-independent systems. However, they can only Tespond accurately to the person or people who trained them, Speaker-independent * Speaker-independent systems do need training by the user. ‘They can recognize speech from anybody according to a general vocabulary, Speaker-independent software is the only real option for businesses, for example—it's not Tealistic to require each customer to train the customer service phone answer system to their voice. Speaker-independent systems are less accurate, but high accuracy limits can sill be attained within processing limits, especially with recent developments in Al and speech recognition technology. Recognition style Speech recognition systems can also be classified depending on the type(s) of utterances (or natural language verbalizations) they can rec ‘ognize, 1) Isolated 2) Connected 3) Continuous Isolated This type of system only understands Separately-spoken words, The speaker must, therefore, pause between each word or command, These systems Srsiganerally setup to Identify words of 0.96 Secha ds or tess.Applications 5.2) Since these are the jest 10 train and program, isolated speech recognition systems are the most common today Connected Connected speech recognition system allow multiple words (or separate utterances) to be yun-together,” but with a minimal pause between them. These systems are set up to identify phrases of 1,92 seconds or less Continuous Continuous recognition systems recognize the natural, conversational speech that we use in our daily lives. Components of Speech Recognition Speech recognition systems today consist of the following key components: Speech capturing device + Speech capturing devices usually include a microphone and an ADC. + A microphone is used to capture speech and convert the sound into electrical signals + The ADC converts the electrical signals into digital data that can be understood by the computer, Digital signal processor A digital signal module processes the raw speech signal, converting the frequency domain and retaining the necessary information only. Preprocessed signal storage {A preprocessed speech signal is kept in the memory of a speech recognition system This helps in carrying out further tasks on speech recognition based on a growing vocabulary base Reference speech patterns Speech recognition systems contain predefined speech patterns already stored in the system's memory that act as a reference for pattern matching. Pattern matching algorithm A special algorithm is used to determine the most likely word(s) by matching the speech pattems to the reference pattern in the speech recognition system's memory. What are the applications of Speech Recognition? Dictation + Dictation is probably the most common use of speech recognition today + Dictation solutions include speech-to-text software that allow the user (© control their devices without typing.Applications _ 5. ip! 23 + They contain some level of computer program that determines what, when and how a robot does something, Difference in Robot System and Other Al Program Here is the difference between the two — Al Programs Robots ] ‘They usually operate in computer-stimulated | They operate in real physical world worlds. The input to an AI program is in symbols | Inputs to robots is analog signal in the form of and rules. speech waveform or images They need general purpose computers to | They need special hardware with sensors and operate on. effectors. Components of a Robot Robots are constructed with the following — + Power Supply- The robots are powered by batteries, solar power, hydraulic, or pneumatic power sources. * Actuators ~ They convert energy into movement. + Electric motors (AC/DC) ~ They are required for rotational movement. + Pneumatic Air Muscles ~ They contract almost 40% when air is sucked in them. * Muscle Wires ~ They contract by 5% when electric current is passed through them. + Piezo Motors and Ultrasonic Motors ~ Best for industrial robots. * Sensors ~ They provide knowledge of real time information on the task environment. Robots are equipped with vision sensors to be to compute the depth in the environment. A tactile sensor imitates the mechanical properties of touch receptors of human fingertips. Computer Vision + This is a technology of Al with which the robots can see. * The computer vision plays vital role in the domains of safety, security, health, access, and entertainment. * Computer vision automatically extracts, analyzes, and comprehends useful information from a single image or an array of images. * This process involves development of algorithms to accomplish automatic visual ‘comprehension. Tasks of Computer Vision * OCR ~ In the domain of computers, Optical Character Reader, a software to convert scanned documents into editable text, which accompanies a scanner.Artificial Intelligence - an * Face Detection ~ Many state-of-the-art cameras come with this re ia enables to read the face and take the picture of that perfect expression. It is used to let a user access the software on correct match. + Object Recognition ~ They are installed in supermarkets, cameras, high-end cars such as BMW, GM, and Volvo. + Estimating Position ~ tis estimating position of an object with respect to camera as in position of tumor in human’s body. Applications of Robotics 52. The robotics has been instrumental in the various domains such as ~ * Industries ~ Robots are used for handling material, cutting, welding, color coating, drilling, polishing, ete. * Military Autonomous robots can reach inaccessible and hazardous zones during war. A robot named Daksh, developed by Defense Research and Development Organization (DRDO), is in function to destroy life-threatening objects safely. * Medicine The robots are capable of carrying out hundreds of clinical tests simultaneously, rehabilitating permanently disabled people, and performing complex surgeries such as brain tumors. + Exploration ~ The robot rock climbers used for space exploration, underwater drones used for ocean exploration are to name a few. + Entertainment ~ Disney's engineers have created hundreds of robots for movie making, PERCEPTIONS + Perception is a process to interpret, acquire, select and then organize the sensory information that is captured from the real world. For example: Human beings have sensory receptors such as touch, taste, smell, sight and hearing. + So, the information received from these receptors is transmitted to human brain to organize the received information, + According to the received information, action is tal F ken by interacting with the environment to manipulate and navigate the objects, * Perception and action are very important concepts in the field of Robotic following figures show the complete autonomous robot, . The{pplications _ 5.2 Fig: Autonomous Robot + Perception presumes sensation, where various types of sensors each converts a ¢ type of simple signal into data of the system. + To put the data together and to make sense out of them is the job of the perception mechanism. + Perception can be seen as a special type of cate recognition) where the inputs are sensory data, judgments and conceptual relations. orization (ot classification, pattern and the outputs are categorical + The difficulty of the task comes from the need of multiple levels of abstraction, where the relations among data items are many-to-many, uncertain, and changing over time, + Perception process of an intelligent system is often (and should be) influenced by internal and external factors beside the signals themselves. Furthermore, perception is not a pure passive process driven by the input. Perception involves interpreting sights, sounds, smells and touch, Action includes the ability to negative through the world and manipulate objects. IF we want to build robots that live in the world, we must understand these processes: Types of Perception Hearing — Speech recognition is the front-end of a system that can perceive and understand spoken language, as used in voice command interface and speech-to-speech translation. Acoustie — phonetic approach Pattern -matching approach Artificial intelligence approach Vision Computational Vision ~ Three primary stages (Stages 1,2 & 3) Acoustic ~ Phonetic Approach + ‘The phonetic units in any spoken language is characterized by set of acoustic properties. + Even though the values varies greatly in these approach the rules governing the variability are straightforward and can be learned easily.§26__Anificial Incetiigence Pattern — Matching Approach + Represents a speech —pattern in the form of a mathematical model * A direct comparison is made between the unknown speeches with possible leareg Pattern during training to identify the unknown, Aritificial Intelligence Approach Attempts to do speech recognition using various Al techniques — Knowledge systems or neural networks. Vision + It begins with a large array of measurement of the light reflected from object surfaces onto the eye. + Analysis then proceed us in multiple stages, with each producing increasingly more ful representations of information in the scene. High-level Perception + The given input data is categorized Low-level perception is bottom-up approach + Itis extremely flexible 10 PLANNING Planning refers to the Process of computing several steps of a problem solving procedure before executing any of them. STRIPS . a L ee as unved ponnist anford Research Institute Problem Solver. + An automated planner. * Planning system for a roboties project: SHAKEY, Developed by Richard Fikes and Nils Nilsson in 1971 at SRI Intemational. + Classical Planning System. + Knowledge Representation: First Order Logic + Algorithm: Forward chaining on rules, STRIPS-Like Planning Formulation + A finite, nonempty set of instances.y ™ + A finite, nonempty set of predicates, w! nary-valued (partial) functions of one of more instances. Each application of a predicate to a specific set of instances is called a positive literal. A logically negated positive literal is called a negative literal. + A finite, nonempty set of operators, each of which has: 1) preconditions, which are positive or negative literals that must hold for the operator to apply, and 2) effects, which are positive or negative literals that are the result of applying the operator. + An initial set which is expressed as a set of positive literals. Negative literals are implied. For any positive literal that does not appear in , its corresponding negative literal is assumed to hold initially. = A goal set which is expressed as a set of both positive and negative literals. STRIPS Planner + STRIPS maintains two additional data structures: © State List - all currently true predicates. «Goal Stack - a push down stack of goals to be solved, with current goal on top of , stack. f + Ifthe current goal is not satisfied by present state, ° Find goal in the add list of an operator, and push operator and preconditions list on stack. (Sub goals). . ‘When a current goal is satisfied, OP it from stack. f uy ‘When an operator is on top of the stack, Midi x « ‘record the application of that operator — update the plan sequence and . use the operator's add and delete lists to update the current state§30__Antificiat intelligence COMPONENTS OF A PLANNING SYSTEM In problem solving systems, it is necessary to perform —, each of the following functions. Choose the best rule to apply next based on the best available heuristic information Apply the chosen rule to compute the new problem state that arises from ijg application. Detect when a solution has been found Detect dead ends so that they can be abandoned and the system’s effort directed jn more fruitful directions. Detect when almost correct solution has been found and employ special techniques to make it totally correct. () Choosing rule to apply The most widely used technique for selecting appropriate rules to apply is first to isolate a set of differences between the desired goal state and the current state and then to identify those rules that are relevant to reducing those differences. If several rules are found, a variety of other heuristic information can be exploited to choose among them. (ii) Applying Rules To handle complex domains, STRIPS problem solver is used. In this approach each operation is described using three li Precondition List —a list of facts which must be true for action to be executed. DELETE list - a list of facts that are no longer true after action is performed; f ADD list - a list of facts made true by executing the action. Basic operations The stack action occurs when the robot arm another object [Y]. The unstack action occurs when the robot arm object Y stack(X,Y): put block X on block Y unstack(X,Y): remove block X from block Y pickup(X): pickup block X from the table putdown(X): put block X on the table f places the object it is holding [X] on top of FORM: STACK(X,Y) f PRE: CLEAR(Y) A HOLDING(X) f ADD: ARMEMPTY() A ON(X,Y) A CLEAR(X) f DEL: CLEAR(Y) A HOLDING(X) f CONSTRAINTS; (X # Y), X# TABLE, Y ¢ TABLE Picks up an object X from on top of_ Applications _ 5.3] + FORM: UNSTACK(X.Y) f + PRE: ON(X.Y) A CLEAR(X) A ARMEMPTY() f + ADD: HOLDING(X) A CLEAR(Y) f + DEL: ON(X.Y) A CLEAR(X) A ARMEMPTY() f + CONSTRAINTS: X # Y, X # TABLE, Y 4# TABLE The pickup action occurs when the arm picks up an object (block) X from the table. f + FORM: PICKUP(X) f «PRE: ONTABLE(X) A CLEAR(X) A ARMEMPTY() f + ADD: HOLDING(X) f * CONSTRAINTS: X # TABLE The putdown action occurs when the arm places the object X onto the table. f + FORM: PUTDOWN(X) f + PRE: HOLDING(X) f + ADD: ONTABLE(X) A ARMEMPTY()A CLEAR(X) f + DEL: HOLDING(X) f + Constraints: X 4 Table (iii) Detecting a Solution {A planning system has succeeded in finding a solution to a problem when it has found a sequence of operators that transform the initial problem state in to goal state. In simple problem solving systems, a straightforward match of the state descriptions has done. (iv) Detecting dead ends As a planning system is searching for a sequence of operators to solve a particular problem, it must be able to detect when it is exploring a path that can never lead to a solution.If the search process is reasoning forward from initial state, it can prune any a path that leads to a state from which the goal state cannot be reached. If the search process is reasoning backward from the goal state it can also terminate 2 path either because it is sure that the initial state cannot be reached or because little progress is being made. In seasoning backward, each goal is decomposed in to sub goals. Each of them in turn may lead to a set of additional sub goals. Sometimes it is easy to detect that there is no way that all sub goals in a given set can be satisfied at once. (») Repairing an almost correct solution5,32 Artificial Intelligence Ive sub problems se The problem is decomposed completely. proceed to solve s ng check that when solutions are combined, they do in fact yield a solution to the original probjg, they do not throw out the solution, look for another one. A slightly better approach to look at the situation that results aang © qoczations corresponding to the proposed solution is executed and to compare he situat desired goal. In most cases, the difference between the two will be smaller “hen between initial state and goal. Now the problem solving can be called again and asked of way of eliminating this new difference. The first solution can then be combined with this se, ‘ne to form a solution to the original problem. An even better way to patch up an almost correct solution is to appeal 10 specific knowledge about what went wrong and then apply a direct path. A Still better way to patch up incomplete solutions is not really to patch them up at butrather to leave them incompletely specified until the last possible moment. Then when a: information as possible is available, complete specification in such a way that no conflicts arise This approach can be thought of as a least commitment strategy. It can be applied in variety of ways. One is to defer deciding on the order in which operations will be performed. GOAL STACK PLANNING One of the earliest techniques to be developed for solving compound goals that may interact was the use of a goal stack This approach was used by STRIPS. A single stack is maintained which contains both goals and operators. Inout sae Conistate ona ‘ ow rastzus . Ox Tanteeys ON PARLE ay samy ON TABLE Wy Fig. 84.4 simple blocks world exemple When we begin solving this problem, the goal stack is simply ON(C,A) “ON(B,D) “ONTABLE(A) “ONTABLE(D) Separate this problem in to sub problems, one for each component of the original go! Two of the sub-problems ONTABLE(A) and ONTABLE(D) are already true in the inital sae OTAD is the abbreviation for ONTABLE(A) * ONTABLE(D) ON(C,A) ON(B.D)a — —_ ONIGA) “ON(B.D) *OTAD ‘at each succeeding step of the problem solving process, the top goal on the stack will be passed. ‘When a sequence of operators that satisfies is found, that sequence 1s applied to the state description, yielding a new description, Applications Next, the goal that is then at the top of the stack is explored and an attempt 1s made to satisfy it, starting from the situation that was produced as a result of satisfying the first goal. This peocess continues until the goal is empty Then as one last check, the original goal is compared to the final state derived from the lication of the chosen operators. If any components of the goal are not satisfied in that state, ss resumed. appli hen those unsolved parts of the goal are reinserted on to the stack and the pro ON(C.A) is replaced with the operator STACK(C.A). STACKIC,A) ON(B.D) ON(C.A) “ON(B.D) “OTAD The preconditions of STACK(C.A) is pushed inside the stack. CLEAR(A) HOLDING(C) CLEAR(A) *HOLDING(C) STACKIC.A) ON(B.D) ON(C.A) “ON(B.D) “OTAD CLEAR(A) is not true. To make CLEAR(A) true, B must be un-stacked from A. This is done by the operator UNSTACK(B.A). UNSTACK(B.A) HOLDING(C) CLEAR(A) “HOLDING(C) STACK(C,A) ON(B.D) ON(C.A)ON(B.D) “OTAD The preconditions of UNSTACK(B.A) are pushed inside the goal stack ON(B.A) CLEAR(B)$34 _Antificial tmrettigence ss ARMEMPTY ON(B,A) °CLEAR(B) “ARMEMPTY UNSTACK(B,A) HOLDING(C) CLEAR(A) *HOLDING(C) STACK(C,A) ON(B,D) ON(C,A) “ON(B,D) *OTAD Here, ON(B,A) is true, CLEAR(B) is true, ARMEMPTY is true. So, they are Popped oy from the goal stack. Thus, all the Preconditions of UNSTACK(B,A) is true. So, UNSTACK(B,4) is placed in the solution list. Solution List={ UNSTACK(B,A) } HOLDING(C) is not true because the arm is holding B. HOLDING(C) CLEAR(A) “HOLDING(C) STACK(C,A) ON(B,D) ON(C,A) *ON(B,D) *OTAD Thus, HOLDING(C) is replaced with the operator PICKUP(C) PICKUP(C) CLEAR(A) *HOLDING(C) STACK(C,A) ON(B,D) ON(C,A) 4ON(B,D) *OTAD The preconditions for PICKUP(C) are pushed inside the goal stack. ONTABLE(C) CLEAR(C) ARMEMPTY “) “CLEAR(C) *ARMEMPTY PICKUP(C) CLEAR(A) “HOLDING(C) STACK(C,A)Applications 5.35 ON(BD) ON(C.A) SON(B,D) SOTAD Here, ONTABLE(C) is true, CLEAR(C) is true, So ONTABLE(C) and CLEAR(C) are popped out trom the goal make the arm empty by s ARMEMPTY ek. ARMEMPTY is not true because the arm is holding C. Thus, cking B on D. This done by the operator STACK(B.D), which replaces STACK(B,D) ONTABLE(C) SCLEAR(C) “ARMEMPTY PICKUP(C) CLEAR(A) HOLDING(C) STACK(C,A) ON(B.D) ON(C,A) SON(B,D) SOTAD. The preconditions of STACK(B,D) is pushed inside the goal stack CLEAR(D) HOLDING(B) CLEAR(D) SHOLDING(B) STACK(B,D) ONTABLE(C) SCLEAR(C) *ARMEMPTY PICKUP(C) CLEAR(A) SHOLDING(C) STACK(C,A) ON(B,D) ON(C,A) *ON(B.D) *OTAD Here, CLEAR(D) and HOLDING(B) are true. This means that the preconditions of STACK(B,D) is true. Thus, CLEAR(D) and HOLDING(B) are popped out and STACK(B,D) 1s placed in the solution list. Solution List { UNSTACK(B,A), STACK(B,D)} ONTABLE(C) “CLEAR(C) “ARMEMPTY PICKUP(C) CLEAR(A) “HOLDING(C) STACK(C,A) ON(B.D)§36__ Artificial Intelligence ON(C,A) “ON(B,D) “OTAD. Now, ONTABLE(C), CLEAR(C) and ARMEMPTY are true. So they are popped out ang PICKUP(C) is placed on the solution list. —. Solution List={ UNSTACK(B,A), STACK(B,D), PICKUP(C)} CLEAR(A) “HOLDING(C) STACK(C,A) ON(B,D) ON(C,A) *ON(B,D) *OTAD Here, CLEAR(A) and HOLDING(C) are true. Thus, CLEAR(A) and HOLDING(C) are Popped out from the goal stack. STACK(C,A) is placed in the solution list, Solution List={ UNSTACK(B,A), STACK(B,D), PICKUP(C), STACK(C,A)} Hierarchical Planning + To solve hard problems, problem solver may have to generate long plans To do this efficiently, eliminate some of the details of the problem until a solution that addresses the main issues is found. Then an attempt can be made to fill in the appropriate details. ABSTRIPS is an approach, which actually planned in a hierarchy of abstraction spaces, in which each of the preconditions at a lower level of abstractions were ignored, (OPERATOR (PRECONDITIONS. (andi...) (oral (w ..)..) (not (exists ...) (06.39) (POSTCONDITIONS: (ADD (..)) (DELETE ( (it (ana (. (DELETE (.....)))) Example: Suppose we want to visita friend in Europe, but you have a limited amount of cash to spend, It makes sense to check air fare first, since finding an affordable flight will be ms difficult part of the task. You should not worry about getting your drive way, planning. route © airport, or parking your car until you are sure you have a flight The ABSTRIPS approach to problem solving is as follows: + First solve the problem completely, considering onl value is the highest possible, ly precond ns whose criticalilYApplications 5.47 mply ignore precondition * Once this is done, use the constructed plan as the outline of a complete plan a consider preconditions at the next lowest criticality level of lower than peak criticality. + Augment the plan with operators that satisfy those preconditions + Again in choosing operators, ignore all preconditions whose criticality level is lew than the level now being considered. + Continue this process of considering less and less critical preconditions until all of the preconditions of the original rules haye been considered. Because this process explores entire plans at one level of detail before it looks at the lower Jevel details of any of them. This is called length-first search, Nonlinear Planning Using Constraint Posting The goal stack planning method attacks problems involving conjoined goals by solving the goals one at a time, in order. A plan generated by this method contains sequence of operators for attaining the first goal, followed by the complete sequence for the second goal. But, the difficult problems cause goal interactions. The operators used to solve one sub problem may interfere with the solution to a previous sub problem. Most sub problems require an intertwined plan in which multiple sub-problems are worked on simultaneously. Such plan is called a nonlinear plan because it is not composed of a linear sequence of complete sub-plans. Let us reconsider the SUSSMAN ANOMALY A ¢ ' bao 2. Site Goat onal oxic.ar* on Taste rar ox. Ds* ON TABLEC!* on ranieuas* ARNEMPTY ON TABLED) ONC. A) ON(A,B)» ONTABLE(A)a ON(B,C) ONTABLE(B)~ ARM EMPTY Achieve ON(B,C) by stacking block B on block C. Complete ON(A,B) by stacking block A on block B$38 _Antificial Intelligence . eee. The initial plan consists of no steps and by studying the goal state a oi the possible Steps are generated. There is no order or detail at this stage. Gradually more detai! 's introduced and constraints about the order of subsets of the steps are introduced until a completely ordereq sequence is created. Tweak Heuristics Using Constraint Posting The Tweak planning method involves the following heuristics. Step Addition — Creating new steps (GPS) for a plan. Promotion - Constraining a step to 20 before another step. Declobberin; Establishment -- Separation — end analysis suggests two steps with end condit operator STACK giving the la Preconditions and followed by its ON(A,B) CLEAR(B) *HOLDING(A) STACK(A,B) ARMEMPTY ON(A,B) CLEAR) HOLDING(A) Many planning methods have introduced heuristic TWEAK planning method brought all these together un introduced/used the following heuristics are mentioned in CLEAR(A) ONTABLE(A) *ARMEMPTY PICKUP(A) ONTABLE(A) —ARMEMPTY HOLDING(A) In this case we need to state that a PICKUP. step. That is to say PICKUP(A) —~ STACK(A,B) ig ~ Placing a new step between two steps to revert a precondition. Simple Assigning a value to a variable to ensure a precondition. Preventing variables being assigned certain values. In this problem means- ions ON(A,B) and ON(B,C) which indicates the yout shown below where the operator is preceded by its ON(B,C) CLEAR(C) *HOLDING(B) STACK(B,C) ARMEMPTY ON(B,C) CLEAR(C) HOLDING(B) S to achieve goals or preconditions. The der one formalism. Other methods that brackets in the following section. CLEAR(C) ONTABLE(B) *ARMEMPTY PICKUP(B) ONTABLE(B) “ARMEMPTY HOLDING(A) Step should precede a corresponding STACK* 7 Applications _ 5.39 PICKUP(B) — STACK(B,C) ‘This gives four steps partially ordered and four unachieved conditions + *CLEAR(A) ~ block A is not clear in initial state. + *CLEAR(B) — although block B is clear in initial state STACK(A,B) with posteondition Remove some unachieved proposition P from S, 4) Achieve P by using step addition, promotion, declobbering, simple establishment, or separation. 4) Review all the steps in the plan including any new steps introduced by step addition to © if any of their preconditions are unachieved. Add to $ the new set of unachieved se preconditions. 5) If's is empty, complete the plan by converting the partial order of steps in to a total onder, and instantiate any variables as necessary, 6) Otherwise go to step 2. Explanation In order for an expert system to be an effective tool, people must be able to interact with it easily. To facilitate this interaction, the expert system must have the following two capabilities in addition to ability to perform its underlying task. Explain its reasoning: In many of the domains in which expert systems operate, people will not accept results unless they have been convinced of the accuracy of the reasoning process that produced those results. This is particularly true, for example, in medicine, where a doctor must accept ultimate responsibility for a diagnosis, even if that diagnosis was arrived at with considerable help from the program. Acquire new knowledge and modification of old knowledge. Since expert systems derive this power from the richness of the knowledge bases they of exploit, it is extremely important that those knowledge bases be as complete and as accurate as possible. But often there exists no standard codification of that knowledge; rather it exists only inside the heads of human experts. One way to get this knowledge in to program is through interaction with human expert, Another way is to have the program learn expert behavior from raw data. Advantage of non-Linear planning Non-linear planning may be an optimal solution with respect to plan length (depending on search strategy used). Disadvantages of Nonlinear planning It takes larger search space, since all possible goal orderings are taken into consideration. Complex algorithm to understand. Algorithm 1) Choose a goal 'g' from the goalset 2) If'g' does not match the state, then Choose an operator 'o' whose add-list matches goal g Push 'o' on the opstack Add the preconditions of 'o' to the goalset. 3) While all preconditions of ‘operator on top of opstack are met in state Pop operator o from top of opstack