1.Introduction to Cognitive computing.pdf

Introduction to Cognitive computing Ms.T.Kavitha Assistant Professor CSE, KITS  Who is smarter? Humans or Computers  God = Natural body(Humans)= Natural Intelligence Natural Intelligence = Ability of humans to make decisions receives a stimulus(Input)...

Introduction to Cognitive computing Ms.T.Kavitha Assistant Professor CSE, KITS  Who is smarter? Humans or Computers  God = Natural body(Humans)= Natural Intelligence Natural Intelligence = Ability of humans to make decisions receives a stimulus(Input) Natural Input = 5 senses – Data(Text, number, images, Signals) These Inputs reaches the brain and helps us to take decision. Humans = Machine(Artificial body) = Artificial Intelligence Vs Structure(hardware) +A.I(software) = Artificial Life. Artificial What is Cognition? Cognition Cognition refers to "the mental action or process of acquiring knowledge and understanding through thought, experience, and the senses". Some of the many different cognitive processes include Thinking, Knowing, Remembering, Judging, And Problem-solving. These are higher-level functions of the brain and encompass language, imagination, perception, and planning.  Cognitive computing is a technology approach that enables humans to collaborate with machines. Based on software and hardware that learns without reprogramming and automates cognitive tasks. Cognitive Cognitive computing as an analog to the human brain, you need to computing analyze in context all types of data, from structured data in databases to unstructured data in text, images, voice, sensors, and video.  Syllabus Introduction to Cognitive computing Content The Uses of Cognitive Systems  Course Objectives: 1. To develop appealing Chabot applications using cognitive computing. 2. To identify and evaluate patterns and complex relationships in large and unstructured data sets 3. Evaluate data in context and presenting relevant findings along with the evidence that justifies the answers Course Outcomes: 1. Outline the importance of cognitive computing Syllabus 2. Analyze the business implications of cognitive computing 3. Apply natural language technologies to business problems 4. Apply machine learning for a specific real world application 5. Develop Chabot applications for business problems 6. Develop Cognitive applications in health care using machine learning  Module 1: Foundations of Cognitive Computing -Cognitive system in education and learning- Cognitive computing as new generation - Uses of cognitive systems-Gaining insights from data-Artificial intelligence-the foundation-Understanding cognition-Understanding complex relationships- The elements of cognitive systems Module 2: Cognitive Computing with Machine Learning - Cognitive Systems in health care –– ML for cancer detection - Design Principles of Cognitive Systems Components of cognitive systems- Building the Corpus- Bringing data into the cognitive system-Machine learning-Hypothesis generation and scoring-Presentation and visualization services Module 3: Cognitive Computing with Inference and Decision Support Systems - Building Chabot application for analyzing customer complaints-Evolution of analytics-core themes-Types of learning-Machine learning algorithms-Cognitive computing model-system-architecture-Cognitive random Modules forest-Cognitive computing system Module 4: Cognitive Computing with Deep Learning - Building Company earning call transcript application-Machine learning Techniques for cognitive decision making – Hypothesis Generation and Scoring- Natural Language Processing - Representing Knowledge - Taxonomies and Ontologies - Deep Learning -Support of Cognitive System Module 5: Cognitive Computing with NLP - Cognitive Assistant for visually impaired using speech Analytics - The role of NLP in a cognitive system, Understanding linguistics, Phonology, morphology, lexical analysis, syntax and syntactic analysis, importance of Hidden Markov models, Semantic Web, Applying natural language technologies to business problems, enhancing shopping experience, fraud detection Module 6: Cognitive Computing with computer vision - Image Recognition for face detection, home room classifier - Introduction to Computer Vision, Computer Vision fundamentals, IBM Watson visual recognition service  TEXT BOOKS: 1. Judith Hurwitz, Marcia Kaufman, Adrian Bowles, “Cognitive Computing and Big Data Analytics”, 1st Edition, Wiley Publisher, 2015, ISBN: 978-1-118-89662-4 2. Vijay V Raghavan, Venkat N. Gudivada, Venu Govindaraju, C.R. Rao Cognitive Computing: Theory and Applications(2016) REFERENCES: Reference 1. Kai Hwang, "Cloud Computing for Machine Learning and Cognitive Applications",MIT press, Cambridge, England, 2017, ISBN:9780262036412 Books: 2. Jerome R. Busemeyer, Peter D. Bruza, “Quantum Models of Cognition and Decision”, Cambridge University Press, 2014. 3. Emmanuel M. Pothos, Andy J. Wills, “Formal Approaches in Categorization”, Cambridge University Press, 2011. 4. Nils J. Nilsson, “The Quest for Artificial Intelligence”, Cambridge University Press, 2009.  The field of Artificial intelligence, Motivation Cognitive science, and Computer science have led to the development of cognitive computing.  1st Era of Computing – Tabulating systems Era of 2nd Era of Computing – Programmable Systems Computing 3rd Era of Computing – Cognitive Computing  Three important concepts help make a system cognitive: Contextual insight from the model, Hypothesis generation (a proposed explanation of a phenomenon), and Continuous learning from data across time. What Makes a A cognitive computing system consists of tools and techniques, including Big Data and analytics, machine learning, Internet of System Things (IoT), Natural Language Processing (NLP), causal induction, Cognitive? probabilistic reasoning, and data visualization. Capability of cognitive computing: To learn, remember, provoke, analyze, and resolve in a manner that is contextually relevant to the organization or to the individual user.  Handwriting Recognition Pattern Recognition Example of Face Detection Self Learning Intelligent Emotion and sentiment analysis Algorithms Systems Vision and Speech Recognition  The ability of the system to understand language to recognize the pattern and to be able to learn from the information to solve more complex challenges Learn—A cognitive system learns. The system leverages data to make inferences about a domain, a topic, a person, or an issue based on training and observations from all varieties, volumes, and velocity of data. Three Model—To learn, the system needs to create a model or fundamental representation of a domain and assumptions that dictate what learning algorithms are used. Understanding the context of how the principles data fits into the model is key to a cognitive system. Generate hypotheses—A cognitive system is probabilistic. A hypothesis is a candidate explanation for some of the data already understood. A cognitive system uses the data to train, test, or score a hypothesis.  1.Privacy Disadvantage 2.Automation Cognitive The amount of new information an individual needs to understand or analyze to make good decisions is overwhelming. Computing as The next generation of solutions combines some traditional a New technology techniques with innovations so that organizations can solve vexing problems. Generation  AI and cognitive computing are expected to contribute nearly 16$ trillion to the global economy by 2030. There will be new uses that emerge that are either focused on The Uses of Horizontal issues (such as security) or Cognitive Industry‐specific problems (such as determining the best way to Systems anticipate retail customer requirements and increase sales, or to diagnose an illness).  Energy: Reduce energy usage by optimizing energy consumption and distribution. Healthcare: Design new drugs and vaccines, diagnose diseases and deliver The Uses of highly personalized medial care. In the healthcare industry, cognitive systems are under development Cognitive that can be used in collaboration with a hospital’s electronic Systems medical records to test for omissions and improve accuracy. The cognitive system can help to teach new physicians medical best practices and improve clinical decision making  Transportation and logistics: Self–driving vehicles, reduce accidents, reduced traffic and major improvements to e-commerce deliveries. City manager to anticipate when traffic will be disrupted by weather events and reroute that traffic to avoid problems. The Uses of Industries: A cognitive system is designed to build a dialog between human and Cognitive machine so that best practices are learned by the system as opposed to Systems being programmed as a set of rules Smart city: Increases our understanding of how to improve the delivery of services to citizens. “Smarter city” applications enable managers to plan the next best action to control pollution, improve the traffic flow, and help fight crime.  Employment: Identify the best candidates for a position and the best positions for a candidate. Smart Homes and Home Robots: Automate and monitor home devices and equipment and provide live-in robot assistants. The Uses of Entertainment and Socialization: Cognitive Identify and recommend experiences and media and help people to find new friends and social circles. Systems Environment: Improve waste processing and recycling and reduce pollution. Business: Automate processes, optimize profits , improve innovation and make better decisions.  As we evolve to cognitive computing we may be required to bring together structured, semi‐structured, and unstructured sources to continuously learn and gain insights from data. Today, much of the data required is text‐based. Natural Language Processing (NLP) techniques are needed to capture the meaning of unstructured text from documents or communications Gaining from the user. NLP is the primary tool to interpret text. Insights from Deep learning tools are required to capture meaning from non- text based sources such as videos and sensor data. Data Visualization is one of the most powerful techniques to make it easier to recognize patterns and understandable in massive and complex data.  NLP is used to identify the semantics of words, phrases, sentences, paragraphs, and other linguistic units in the documents and other Natural unstructured data found in the corpus. Language One important use of NLP in cognitive systems is to identify the statistical patterns and provide the linkages in data elements so that Processing the meaning of unstructured data can be interpreted in the right context.  In a probabilistic system, there may be a variety of answers, depending on circumstances or context and the confidence level or probability based on the system’s current knowledge. A deterministic system would have to return a single answer based on the evidence, or no answer if there were a condition of Domains uncertainty. Where The cognitive solution is best suited to help when the domain is complex and conclusions depend on who is asking the question and Cognitive the complexity of the data. Computing Is For example, in the medical diagnostic example, the cognitive system may ask the physician to perform additional tests to rule out Well Suited or to choose certain diagnoses.  Although the seeds of artificial intelligence go back at least 300 years, the evolution over the past 50 years has had the most impact Artificial for cognitive computing. Artificial Narrow intelligence Intelligence as Artificial General Intelligence the Foundation Artificial Super Intelligence of Cognitive As computer science evolved, computer scientists assumed that it would be possible to translate complex thinking into binary coding Computing so that machines could be made to think like humans. Alan Turing’s Alan Turing, a British mathematician whose work on cryptography was recognized by Winston Churchill as critical to victory in WWII, was also view on AI as a pioneer in computer science. Turing turned his attention to machine learning in the 1940s. the Foundation “Computing Machinery and Intelligence” of Cognitive ” Turing argued that with advancement in digital computing, it would be possible to have a learning machine whose internal processes were Computing unknown, or a black box. The test consisted of two humans and a third person that inputted questions for the two people via a typewriter. The goal of the game was to determine if the game players could determine which of the three participants was a human and which was a “typewriter” or a computer. In other words, the game consisted of human/machine interactions. He was making the distinction between the ability of the human to intuitively operate in a complex world and how well a machine can mimic those attributes. Norbert Another important innovator was Norbert Weiner, whose 1948 Weiner’s view book, “Cybernetics or Control and Communication in the on AI as the Animal and the Machine”, defined the field of cybernetics(the science of communications and automatic control systems in both Foundation machines and living things). He studied the continuous feedback that occurred between a guided of Cognitive missile system and its environment. Computing Weiner’s theories on the relationship between intelligent behavior and feedback mechanisms led him to determine that machines could simulate human feedback mechanisms. Arthur Lee Samuel’s view Arthur Lee Samuel is credited with developing the first on AI as the self‐learning program for playing checkers. Foundation Checkers is a classic board game, dating back to around 3000 BC. of Cognitive Two machine‐learning procedures have been investigated in some detail using the game of checkers. Computing Enough work has been done to verify the fact that a computer can be programmed so that it will learn to play a better game of checkers than can be played by the person who wrote the program. Herbert Simon’s view Herbert Simon, who won the Nobel Prize for Economics in 1978, had an ongoing interest in human cognition and decision making on AI as the that factored into all his research. Foundation Simon and his colleagues such as Alan Newell assumed that a simple adaptive mechanism would allow intelligence to be of Cognitive captured to create an intelligent machine. Computing Simon laid out the concept of natural language processing and the capability of computers to mimic vision. He predicted that computers would play chess at the grand master level. Feigenbaum’s In 1965, after joining the computer science faculty at Stanford University, Feigenbaum and Nobel laureate Joshua Lederberg view on AI as started the DENDRAL project, which was later referred to as the first expert system. the Foundation Feigenbaum said that the DENDRAL project was important of Cognitive because it showed that “The dream of a really intelligent machine was possible”. Computing Today, expert systems are used in the military and in industries such as manufacturing and healthcare. DARPA In the late 1980s DARPA sponsored the FORCES project, which was part of the Air Land Battle Management Program. This was an sponsored the expert system designed to help field personnel make decisions based on historical best practices. FORCES Although military-based research continued to be funded by project DARPA, commercial based funding. The area of machine Subfields of AI including machine learning, ontologies, rules learning management, pattern matching, and NLP continued to find their way into a myriad of products over the years. Today, most of the focus is on the area of machine learning algorithms that provide a mechanism to allow computers to process data in a methodical way. But much of the focus of machine learning is dealing with ambiguity because most data is unstructured and open to many different interpretations.  The word cognition, from the Latin root gnosis, meaning to know and learn, dates back to the 15th century. Understanding how the human brain works and processes Understanding information provides a blueprint for the approach to cognitive computing. Cognition By understanding cognition we can build systems that have many of the characteristics required to continuously learn and adapt to new information. Two Cognitive science—The science of the mind. disciplines of Computer science—The scientific and practical approach to Cognitive computation and its applications. It is the systematic technique for translating this theory into practice. Computing:  The main branches of cognitive science are Cognitive Psychology and science Neurology  Psychology - primarily an applied science, in helping diagnose and treat mental/behavioral conditions Psychology  Neurology - primarily applied, in diagnosis/treatment of neurological conditions. A cognitive architecture is also directly tied to how the neurons in the brain carry out specific tasks, absorb new inputs dynamically, and understand context. Neurology  Cognitive scientists, in studying the human mind, have come to understand that human cognition is an interlinking system of systems that allows for information to be received from outside inputs, which is then stored, retrieved, transformed, and transmitted. Cognitive Likewise, the development of the computer field has accelerated the science field of cognitive sciences. A foundational principle of cognitive science is that an intelligent system consists of a number of specialized processes and services that interact with each other. Cognitive There may be an architecture that is related to human senses such as seeing, understanding speech, and reacting to tastes, smells, and science to touch. cognitive The human brain is architected to deal with the mental processes of perception, memory, judgment, and learning. computing  The human may have a bias that leads to conclusions that are erroneous. For example, the human may look at one research study that states that there are some medical benefits to chocolate and conclude that Machines do eating a lot of candy will be a good thing. not have bias In contrast, a cognitive architecture will not make the mistake of assuming that one study or one conclusion has an overwhelming relevance unless there is actual evidence to draw conclusions. Unlike humans, machines do not have bias unless that bias is programmed into the system.  Traditional architectures rely on humans to interpret processes into code. AI assumes that computers can replace the thinking process of The process of humans. cognitive With cognitive computing, the human leverages the unique ability of computers to process, manage, and associate information to computing expand what is possible.  It is quite complicated to translate the complexity of human thought and actions into systems. In human systems, we are often influenced by emotion, instinct, habits, and subconscious assumptions about the world. Cognition is a foundational approach that leverages not just how we Complexity of think, but also how we act and how we make decisions. understanding Example human Why does one doctor recommend one treatment whereas another doctor recommends a completely different approach to the same disease? Why do two people raised in the same household with a similar experience grow up to have diametrically opposed views of the world?  The two systems of thought and how they relate to how cognitive computing works. Two Systems System 1(Intuitive thinking) thinking is the type of intuitive reasoning that can be analogous to the type of processing that can be of Judgment easily automated. and Choice System 2(Controlled and rule‐centric thinking) thinking is the way we process data based on our experiences and input from many data sources (thinking is related to the complexities of cognitive computing).  Thinking begins almost from the moment we are born. We learn to see objects and understand their relationships to System ourselves. 1—Automatic For example, we associate our mother’s voice with safety. We associate a loud noise with danger. These associations form the Thinking: basis of how we experience the world. The child with a cruel Intuition and mother will not have the same association with the mother’s voice as the child with the kind mother. Of course, there are other issues Biases at play as well. The child with a kind mother may have an underlying mental illness that causes irrational actions. An average child who associates a loud noise with fun may not feel in danger.  The chess protégée who becomes a master automatically learns to make the right moves. The chess master not only knows what his next move should be but also can anticipate what move his opponent will do next. That chess master can play an entire game in his mind without even System touching the chessboard. 1—Automatic Likewise, emotions and attitudes about the world are automatic, as well. Thinking: Example: Intuition and If a person is raised in a dangerous area of a city, he will have Biases automatic attitudes about those people around him. Those attitudes are not something that he even thinks about and cannot easily be controlled. These attitudes are simply part of who he is and how he has assimilated his environment and experiences.  The benefit of System 1 thinking is that we can take in data from the world around us and discover the connections between events. It Advantage and is easy to see that System 1 is important to cognitive computing because it allows us as humans to use sparse information we collect disadvantage about events and observations and come to rapid conclusions. of System 1 can generate predictions by matching these observations. Disadvantage: However, this type of intuitive thinking can also be System 1 inaccurate and prone to error if it is not checked and monitored.  System 2 thinking observes and tests assumptions and observations, instead of jumping to a conclusion based on what is assumed. Example System A potential cancer treatment seems promising. All the preliminary data indicates that the drug will eradicate the cancer cells. However, the 2—Controlled, treatment is so toxic that it also destroys healthy cells. Rule‐Centric, System 1 thinking would assume that the fact that cancer cells are destroyed is enough to determine that the drug should immediately be put on the and market. However, System 1 thinking often includes bias. Although it may Concentrated appear that an approach makes sense, the definition of the problem may be ill‐defined. Effort System 2 thinking slows down the evaluation process and looks at the full context of the problem, collects more data across silos, and comes up with a solution.  Advantage:. Because System 2 is anchored in data and models, it takes into account those biases and provides a better outcome. Predicting outcomes is a complex business issue because so many factors can change outcomes. This is why it is important to combine Advantages of intuitive thinking with computational models. System 2 The ability to analyze massive amounts of information related to the problem being addressed and to reason in a deliberate manner.  Combining System 1 intuitive thinking with System 2 deep analysis is critical for cognitive computing. Combining System 1 and System 2  Unstructured text‐based information sources have to be parsed so that it is clear what content is the proper nouns, verbs, and objects. This process of categorization is necessary so that the data can be Understanding consistently managed. Data from unstructured sources such as images, video, and voice have to be analyzed through deep analytics Complex of patterns and outliers. Relationships For example, recognition of human facial images may be facilitated by analyzing the edge of the image and identifying for Between patterns that can be interpreted as objects such as a nose versus an eye. Systems It is most helpful to have an approach that is highly interdisciplinary and provides a framework to help individuals find answers to some fundamental questions based on continually refining the elements of the information sources that are most relevant.  Cognitive systems are intended to address real‐world problems in an adaptive manner. This adaptive systems approach is intended to deliver relevant data‐driven insights to decision makers based on advanced analysis of the data. Adaptive For example, the system could be looking at the stock market and Systems the complex set of information about individual companies, statistics about performance of economies, and competitive environments. The goal of the adaptive system would be to bring these elements together so that the consumer of that system gains a holistic view of the relationship between factors.  A cognitive system consists of many different elements, ranging from the hardware and deployment models to machine learning and applications The Elements of a Cognitive System  In addition, these applications may need to infuse processes to gain insight about a complex area such as preventive maintenance or treatment for a complex disease. An application may be designed to simulate the smartest customer Cognitive service agent. Applications A well‐designed cognitive system provides the user with contextual insights based on role, the process, and the customer issue they are solving. The solution should provide the users insights so they make better decisions based on data that exists but is not easily accessible.  Visualization services help to communicate results by providing a way to demonstrate the relationships between data. Presentation The two basic types of data visualizations are static and dynamic. and Visualization may depend on color, location, and proximity. Visualization Other critical issues that impact visualization include shape, size, and motion. Services Making this data interactive through a visualization interface can help a cognitive system be more accessible and usable.  Machine learning is the technique that provides the capability for the data to learn without being explicitly programmed. Cognitive systems are not static. Continuous Rather, models are continuously updated based on new data, Machine analysis, and interactions. A machine learning process has two key elements: Learning Hypothesis generation and Hypothesis evaluation  Industries ranging from transportation to healthcare use sensor data The Learning to monitor speed, performance, failure rates, and other metrics and then capture and analyze this data in real time to predict behavior Process and change outcomes. The Corpus, A corpus is the knowledge base of ingested data and is used to manage codified knowledge. Taxonomies, A taxonomy provides context within the ontology. Ontologies are and Data often developed by industry groups to classify industry‐specific elements such as standard chemical compounds, machine parts, or Catalogs medical diseases and treatments.  Data analytics services are the techniques used to gain an understanding of the data ingested and managed within the corpus. Data Analytics Typically, users can take advantage of structured, unstructured, and Services semi‐structured data that has been ingested and begin to use sophisticated algorithms to predict outcomes, discover patterns, or determine next best actions.  Because cognitive computing centers around data, it is not surprising that the sourcing, accessing, and management of data Data Access, play a central role. Metadata, and In a cognitive system these data sources are not static. There will be a variety of internal and external data sources that will be included Management in the corpus. Services Therefore, as in a traditional system, data has to be vetted, cleansed, and monitored for accuracy. Infrastructure An organizations can leverage Software as a Service (SaaS) and applications and services to meet industry‐specific requirements. Deployment A highly parallelized and distributed environment, including compute and storage cloud services, must be supported. Modalities