MSYS20 Notes.pdf

MODULE 1: Introduction Systems Thinking involves To Information Systems 1) Recognizing Interconnections - What are the key connections between INTRODUCTION TO...

MODULE 1: Introduction Systems Thinking involves To Information Systems 1) Recognizing Interconnections - What are the key connections between INTRODUCTION TO system parts? INFORMATION SYSTEMS 1) Systems Thinking 2) Information Systems Definition What is a SYSTEM? An orderly grouping of interdependent components linked together according to a plan to achieve a specific objective. Defining a System Characteristics: 2) Identifying and Understanding 1) Organization Feedback 2) Interaction - What are the feedback 3) Interdependence loops that affect system 4) Integration behavior? 5) Central Objective Elements of a System: 1) Outputs and Inputs 2) Processors 3) Feedback 4) Environment 5) Boundaries and Interfaces Types of System: 1) Physical or Abstract Systems 2) Open or Closed Systems 3) Man-made Information Systems 3) Understanding Systems Structure [ - How are the system elements organized? 4) Identifying and Understanding Dynamic Behavior - What subject matter expertise is needed to understand the nature of the system? 5) Reducing Complexity by - How does the system Modeling Systems work? What are the Conceptually sequences of activities? - What is my view of the What are the flows system? What lens can I (inputs and outputs)? use to create a - What are the resources representation of system? used? When are they used? What levels of these resources are expected (as inputs? as outputs?) How does the system behaves given variations in these resources?  6) Understanding Systems at Different Scales - What’s the big picture? - What are the subsystems that make up the system? What is an INFORMATION SYSTEM? Systems Thinking Everything can be considered a system Definition ○ Composed of interrelated components Information system (IS) is the ○ Designed to meet a study of complementary certain goal networks of hardware and ○ Relates with other software that people and “systems” organizations use to collect, ○ Each component can be filter, process, create and considered “subsystem” distribute data Seeing interrelationships Information systems are combinations of hardware, software and telecommunications networks that people build and use to collect, create and distribute useful data, typically in organizational settings Information systems are interrelated components working together to collect, process, store and disseminate  information to support decision Data and Information making, coordination, control, analysis, and visualization in an Systems organization. The goal of an Information System is to transform data into information in Information System order to generate knowledge that can be used for decision making A set of interrelated components that collect, manipulate, store and disseminate data and information and provide a feedback mechanism to meet an objective Accessible - must be easily accessible by users to meet their Information Systems in needs in the right format at the right time. Access should be Action secure and prevent Here are some examples of IS that you unauthorized access. may or may not have noticed in your Consistent - contains no everyday life discrepancies, and the same Project Noah - Disaster and Risk measurements or structure are Management used regardless of the data FASSSTER - Public Health and source. Epidemiology Complete - Contains all AISIS - Education information and is not missing anything Timely - Delivered when needed Accurate - Free of errors and bias, and can be verified or validated Cost - Balance the value of information to the cost of producing it Relevant - Important to decision makers  Clear - Not overly complex, easy Educational background on to understand business is relevant Chief Privacy Officer Components of an Handles privacy issues and helps shape policy about Information System sensitive and confidential data Chief Security Officer Ensures that confidential information is protected from hackers, disasters, accidents and rogue employees Systems Administrator Installs, manages and updates servers Enterprise Systems Officer 1) People Develops, installs, maintains, People can be the most oversees the organization’s important element in most mission-critical software computer-based information applications systems. They make the Data Center Operations difference between success and Maintains environmentally failure for most organizations. controlled areas in which servers, mainframes and communications equipment are located Helpdesk Provides service to internal and external customers on technology issues 2) Technology Chief Information Officer Sometimes called “VP of Information Systems” Reports directly to CEOs or other VPs (usually under finance or administration) Great leadership abilities, communication skills, knowledgeable on ICT  3) Process Information system... A business process is a set of is a set of interrelated hardware, activities designed to achieve a software, networks, data, processes task, and organizations and people, working together to implement information systems collect, manipulate, store and support, streamline and disseminate data and information, and sometimes eliminate business provide a feedback mechanism processes to meet an objective. 4) Data Raw ingredient for every information system Converted into digital format, integrated and shared across systems Stored in databases SUMMARY Systems Thinking involves… 1) Recognizing Interconnections 2) Identifying and Understanding Feedback 3) Understanding Systems Structure 4) Identifying and Understanding Dynamic Behavior 5) Reducing Complexity by Modeling Systems Conceptually 6) Understanding Systems at Different Scales Module 2: Information Laptop computer Mobile phones Technology Tablet computer Server Infrastructure Storage devices Input and Output devices A. Hardware and Software Digital Devices in everyday Technology objects Automobiles/cars Application of scientific Refrigerators knowledge for practical Washing Machines purposes Smart Watches Technological components of information system: hardware, software, data, network Digital devices process electronic signals into discrete values Computers use digital electric signals, expressing all information using only ‘0’ and ‘1’ Hardware Computing device and its parts Tangible part of the IT Infrastructure that handles Digital Data ○ Collects Data ○ Processes Data ○ Stores Data ○ Produces User Output Digital Devices Desktop computer  Storage Permanent Storage - Does not lose its data when unpowered Hard Disk Drives Used for long term data storage; non-volatile storage Solid State Drives Instead of spinning disks, it uses flash memory using EEPROM chips, which is much faster Removable Media Floppy disks -> CD-ROM -> USB Computer System Concept Non-permanent Storage - Loses its data when unpowered - Faster data access compared to permanent storage Random-Access Memory Stores all the data and instructions that needs to be processed Faster than a hard disk Can be thought of as the computer’s short-term memory Central Processing Unit CPU Cache “Brain” of the device Reserved high speed access Carries commands sent to it by area within RAM the software and returns results For memory blocks that are to be acted upon most frequently needed by the Hertz - speed of a CPU; 1 hertz is CPU one cycle per second CPU chips can contain multiple Input and Output Devices processors e.g. dual-core, quad-core Input Intel Core i7 - up to 20 cores Keyboard Intel Core i9 - up to 16 cores Mouse Scanner Microphone Camera/Webcams  Touch screens Output Display Monitor Printer Speaker Network Connection Network Interface Cards (NIC) - Ethernet Network Port Speed of your computer Compo Speed Units Descri nent measur ption Wireless Adapters ed by CPU Clock GHz Hertz speed (billions indicat of es the cycles) time it takes to comple How do we put them all te a cycle. together? Mother Bus MHz The MOTHERBOARD - a piece of board speed speed hardware that connects all the at components (CPU, video cards, which sound cards, RAM, hard devices data etc.) together can move across the bus RAM Data Mb/s The transfer (million time it  Application software - performs rate s of takes bytes for data specific tasks and functions per to be (productivity, programming, second) transfer enterprise system, etc.) red from memor y to the system measur ed in Megab ytes. Hard Access ms The Disk Time (millise time it cond) takes for the drive to locate Operating System the 1) Integrated system of programs data to that… be accesse a) Manages the operations d. of the CPU b) Controls the input/output, Data MBit/s The storage resources, and transfer time it activities of the computer rate takes for data system to be c) Provides support services transfer as the computer executes red application programs from 2) The operating system must be disk to loaded and activated before system other tasks can be accomplished Software The set of instructions that tells the hardware what to do Operating System - Manage the hardware and create an interface between the hardware and user  Programming Software - environment where developers can write codes and convert it into programs or software that run on a computer Applications for the Enterprise Tools for managing company Productivity Software operations and business decision making Tools for the workplace Enterprise Resource Planning ○ System that manages and bring together an entire organization’s operations Customer Relationship Utility Software and Management ○ Manages a company’s Programming Software customers, marketing Utility software - programs that allow and sales activities you to fix or modify your computer, Supply Chain Management e.g., anti-virus, anti-malware ○ Handles the interconnection between the different stages of product development and delivery Supplier Relationship Management ○ Manages a company’s suppliers, negotiations and supplier contracts Human Capital Management  ○ Manages a company’s ○ Not consumed when employee data and HR used operations ○ Can be stolen without being gone and can be used by many people at the same time ○ Unique to the organization Information Processed data possess context, relevance and purpose Involves manipulation of raw data to obtain insights, trends, B. Data and Networks patterns to be used in decision making Data Information that facilitates action is Knowledge Information Requirements Combining knowledge and in Decision Making Levels experience to produce a deeper understanding of a topic is Wisdom. Data, Information, Knowledge, Wisdom Data and Information Data Represents facts whether digital or written Digital Data is an asset with unique properties ○ Not tangible but durable ○ Easy to copy and transport but difficult to reproduce when destroyed Organizing Business Data relational design. It could be a tree, a graph, a network, or a and Information key-value pairing Transactions and Decision Hierarchical Databases Making Oldest and most rigid database model, used in early mainframe DBMS Data is organized into a tree-like structure with mandatory parent/child relationships Each parent can have many children, but each child has only one parent Databases Any collection of stored data, regardless of structure or content An organized collection of related data with the purpose of putting data into context and providing tools for aggregation and analysis Common challenges for not using a database: Relational Databases ○ No control of redundant Commonly used in business data organizations with a few ○ Violation of data integrity exceptions ○ Relying on human Organized in tables, where memory to store and to tables in the database are sets search needed data of relations with identical structure Types of Databases Each table has a seat of fields which define the structure of Non-Relational Databases the data Stores data as simple strings or Row-oriented, where a row is a complete files record or an instance of a set of The primary differentiating fields in a table factor is the storage structure A field may be identified as a itself, where the data structure is primary key, which uniquely no longer bound to a tabular identifies a record in a table  Networking and Communication Components of a Simple Computer Network Database Management Network Operating System - System (DBMS) routes and manages communications on the Application to create and network resources; it can also manage a database, change a reside on every computer database’s structure or design, Switch - a network device that create queries, develop reports connects multiple hosts and do analysis together and forwards packets Other uses of a DBMS based on their destination ○ Data Security: within the local area network create/manage user (LAN) access to the database Router - a device that receives ○ Data Backups: ensure and analyzes packets and then availability of database by routes them towards their creating/managing destination database backups Packet - the fundamental unit of data transmitted over a network Understanding the Data Computer Systems - servers, Subsystem clients (PCs, mobile devices, digital gadgets) that contain, sends and receives the data and applications that need to be shared Network Processors ○ Switch - a network device that connect multiple hosts together and forwards packets based on their destination within the local area network (LAN) ○ Router - a device that receives and analyzes packets and then routes  them towards their waves can be picked up of up to destination approx. 65ft. By devices with Network Media - wired or wireless adapter wireless Network Protocols and Standards - sets rules and guidelines on the transmission technologies and data ○ TCP/IP ○ CDMA, GSM, LTE, 5G Other Scope of Business Corporate Network Networks Infrastructure Intranet - provides web-based resources from users within the Today’s corporate network organization; available only thru infrastructure is a collection of LAN; e.g. employee portal many networks from public Extranet - part of the company’s switched telephone network, to network that can be made the internet, to corporate local available securely to those area network (LAN) linking outside the company; access workgroups, departments, or can be provide via virtual private office floors network (VPN) LAN - designed to connect computers and digital devices within a 500m radius. Using ethernet cable or WiFi. Ethernet - a way of connecting computers of network devices in a physical space (wired network) Wi-Fi access point - takes a network signal and converts it into radio waves and these radio What is the Internet? The world’s most extensive public communication network Internet Service Providers (ISP) - commercial organization with a permanent connection with the internet providing temporary connection to subscribers (homes/ businesses) The internet addressing is based on TCP/IP networking protocol ○ TCP - Transmission Control Protocol handles the movement of data between computers ○ IP - Internet protocol is Internet is based on a responsible for the Client/ Server Technology delivery of packets; each computer on the internet Client: PCs or computing has an IP address, e..g. devices are networked and www.microsoft.com is connected to servers 207.46.250.119 Server: Provides client computers with variety of services and capabilities Web/Application Server: A web server will serve a web page to a client in response to a request for service. Application server software handles all application operations between a user and Domain Name System organization. (DNS) Used to remember internet addresses easier Convert domain names to IP addresses Domain Name - English-like name that corresponds to an IP Address Other Communication hosts together and forwards packets based Tools on their destination VOIP (Voice over IP) allows within the local area analog signals to be converted network (LAN) to digital signals ○ Router - a device that receives and analyzed packets and then routes them towards their destination Network Media - wired or wireless Software - system software found in the computers and Bluetooth - standard method of network processors specialized wirelessly connecting nearby for transmission of data; devices within approx. 300ft. application software used by end-users in the network environment Type of network - LAN, WAN, Intranet, Extranet, Internet Network Protocols and Standards – sets rules and guidelines on the transmission technologies and data ○ TCP/IP ○ CDMA, GSM, LTE, 5G Understanding the Network Subsystem Computer Systems - servers clients (PCs, mobile devices, digital gadgets) that contain, sends and receive the data and applications that need to be shared Network Processors ○ Switch - a network device that connects multiple C. Trends in Global Information Use of Internet is growing all over the world Technology RECAP: Technological components of IS Social In 2024, there are 5.0B social media users worldwide This number is expected to jump to approximately 5.85 Information Technology billion users by 2027 The most used social media Trends platform in the world is 1) Global Digitalization Facebook, with 2.9 million a) Internet monthly active users across the b) Web 2.0 world c) Social The average person spends d) Mobile about 145 minutes on social 2) Technology Convergence media every day a) Cloud Computing The most common way people b) Internet of Things access social media is a mobile c) Data Science device d) Artificial Intelligence Mobile Websites became interactive, and user experience became customizable Defining features of Web 2.0 ○ Interactivity ○ Real-time user control ○ Social sharing ○ User generated content Technologies and services behind these features ○ Cloud Computing ○ Social Networks ○ Blogs ○ Apps Mobile Applications Web 2.0 Web 2.0 is a second-generation interactive internet-based service We are at the age of the Internet of Things Data Science Fourth Industrial Artificial Intelligence Revolution Characterized by Global Digitalization Cloud Computing “Cloud” refers to applications, services, and data storage located on the internet ○ Changed how we can access and use software and hardware resources ○ For businesses: changed acquisition, financing practices ○ Software-as-a-Service Microsoft 365 vs Microsoft Office ○ Platform-as-aService Hosted development via the Internet Technology Convergence ○ Infrastructure-as-a-Servic e Understanding Emerging Subscribe vs buy Technologies and maintain Convergence of servers Information Technologies Cloud computing Internet of Things The Internet of things, IoT is a network Big Data: Business of interrelated devices that connect and exchange data with other IoT Intelligence, Analytics, devices and the cloud Data Science IoT devices are typically embedded with technology such as sensors and software and can include mechanical and digital machines and consumer objects  Business Intelligence - infrastructure (connected technologies) for data warehousing, integration, reporting and analyzing data Examples coming from the business environment Data Analytics - examining large data sets to identify trends, develop charts and create visual presentations to help business make more strategic decisions Data Science - design and constructing new processes for data modeling and production using prototypes, algorithms, predictive models and custom analysis to help solve complex problems  able to surpass the knowledge and capabilities of humans Reactive Machine AI: AI capable of responding to external stimuli in real time; unable to build memory or store information for future Limited Memory AI: AI that can store knowledge and use it to learn and train for future tasks Artificial Intelligence Theory of Mind AI: AI that can Consists of computer-based sense and respond to human systems (both hardware and emotions, plus perform the software) that attempt to tasks of limited memory emulate human behavior machines Such systems would be able to Self-Aware AI: AI that can ○ Learn languages, recognize others’ emotions, plus accomplish physical tasks has a sense of self and (robotics) human-level intelligence; the ○ Use a perceptual final stage of AI. apparatus that informs Examples physical behavior and language ○ Emulate human expertise and decision making Some Types of AI Weak AI - also known as narrow AI, is designed and trained to complete a specific task Strong AI - also known as artificial intelligence (AGI), describes programming that can replicate the cognitive abilities of the human brain; uses fuzzy logic to apply knowledge from one domain to another and find a solution autonomously Super AI - also known as artificial superintelligence (ASI), AI Technology Malware: Malicious Components Software A computer virus is a rogue software program that attached itself to other software programs or datafiles to be executed, usually without user knowledge Worms are independent computer programs that copy themselves from one computer to other computers over a network Trojan horse is a software program that appears to be benign but then does something other than expected. The Trojan horse is not itself a virus because it does not replicate, but it is often a way for viruses or other malicious code to be introduced into a computer system D. Information Security Ransomware tries to extort Security Threats money from users by taking control of their computers or Security Challenges and displaying annoying pop-up Vulnerabilities messages Spyware are small programs All components of Information that install themselves Systems (IS) are vulnerable to security surreptitiously on computers to threats and challenges monitor user web-surfing activity and serve up advertising Keyloggers record every keystroke made on a computer to steal serial numbers for software, to launch Internet attacks, to gain access to email accounts, to obtain passwords to protected computer systems, or to pick up personal  information such as credit card In a Denial-of-service (DoS) or bank account numbers attack, hackers flood a network server or web server with many thousands of false communications or requests for services to crash the network. The network receives so many queries that it cannot keep up with them and is thus unavailable to service legitimate requests Identity Theft Identity theft is a crime in which an imposter obtains key pieces of personal information, such as social security numbers, driver’s license numbers, or credit card numbers, to impersonate someone else Hackers Phishing involves setting up Hacker is an individual who fake websites or sending email intends to gain unauthorized messages that look like those of access to a computer system legitimate businesses to ask Cracker is typically used to users for confidential personal denote a hacker with criminal data. The email message intent instructs recipients to update or Spoofing involves redirecting a confirm records by providing web link to an address different social security numbers, bank from the intended one, with the and credit card information, and site masquerading as the other confidential data either by intended destination responding to the email Sniffer is a type of message, by entering the eavesdropping program that information at bogus website or monitors information traveling by calling a telephone number over a network. They enable Click fraud occurs when an hackers to steal proprietary individual or computer program information form anywhere on a fraudulently clicks an online ad network, including e-mail without any intention of messages, company files and learning more about the confidential reports advertiser or making a purchase Internal Threat Authentication Types of Insider Threats Ensures that the person accessing the information is, Negligent - Insiders who pose indeed, who they represent an unintentional threat due to themselves to be human error and lack of security awareness Malicious - Current or former employees who abused their access to steal intellectual property for personal gains Third-Party - Vendors who misuse their access and compromise the security of critical data Information Security Access Control The Information Security Ensures that users can only Triad access the information resources and perform system Confidentiality - restrict access tasks that appropriate to those who are allowed to see it; NTK (need to know) Integrity - information truly represents its intended meaning; it has not been altered Availability - information can be accessed and modified by authorized personnel in an appropriate timeframe Encryption Process of encoding data upon its transmission or storage so that only authorized individuals can read it  Physical Security Proper Datacenter Operation and Security Firewall and Intrusion Backups Detection Systems A comprehensive backup plan is Firewall essential for information security A hardware firewall is a device ○ Full understanding of the that is connected to the organization’s network and filters packets information resources based on a set of rules ○ Regular backups for all Software firewall runs on the data operating systems and ○ Offsite storage of backup intercepts packets as they arrive data sets to the computer ○ Test of data restoration IDS Provides the capability to identify if the network is being attacked  over their own IT infrastructure use these private clouds. VPN VPN is a secure, encrypted, private network that has been configured within a public network to take advantage of the economies of scale and management facilities of large Personal Information networks, such as the Internet Security Provides your firm with secure, encrypted communications at a 1) Keep your software up to date much lower cost 2) Install antivirus software and Also provide a network keep it up to date infrastructure for combining 3) Be smart about your voice and data networks connections 4) Backup your data 5) Secure accounts with two-factor authentication 6) Make your passwords long, strong and unique 7) Be suspicious of strange links and attachments Private Cloud A private cloud is operated solely for an organization. It may be managed by the organization or a third party and may be hosted either internally or externally. Companies that want flexible IT resources and a cloud service model while retaining control CISCO Introduction to What is the difference between AI and augmented Artificial Intelligence intelligence? Artificial intelligence: When learning about artificial intelligence, you’ll come across the Recommends merchandise you term augmented intelligence. Both might like to buy on the internet terms share the same objective, but Alerts you if your smartwatch or have different approaches. fitness band detects low oxygen Augmented intelligence has a in your bloodstream, modest goal of helping humans with inflammation, or an unhealthy tasks that are not practical to do. For spike in blood sugar example, “reading” 1000 pages in an Scans your social media posts to hour. In contrast, artificial intelligence learn more about what you are has a lofty goal of mimicking human thinking thinking and processes. However, it’s Helps banks invest money in important to note that AI today is not your family’s bank accounts to mature enough to perform keep the economy around you independent tasks such as diagnosing growing cancer. Module 1: What is Artificial Example Human Intelligence - operate vehicle Intelligence? Artificial intelligence - self driving What is AI? feature (replaces need of human) Augmented intelligence - collision AI plays an often invisible role in detection; blindspot avoidance ( everyday life, powering search engines, machines and human both working product recommendations, and together) eg. screen reader for blind speech recognition systems. Artificial intelligence (AI) refers to the Machines Humans ability of a machine to learn patterns Ingesting Generalizing and make predictions. AI does not Data Information replace human decisions; instead, AI Repetitive Creative adds value to human judgment. Accurate Emotional In its simplest form, artificial So, what continues to drive the intelligence is a field that combines development of AI? computer science and robust datasets to enable problem-solving. As computing power and algorithms become more powerful and data volumes increase, companies will Example: Sorting Meat with very little adopt new use cases for AI time technologies. Companies will embed Artificial Intelligence smart systems into their applications to drive innovation and efficiencies, enhance employee experience, automate tasks, decrease costs, and improve revenue. What does AI do? Artificial intelligence machines (researchers call them “AI services”) don’t think. They calculate. They represent some of the newest, most sophisticated calculating machines in human history. Some can perform what’s called machine learning as they acquire new data. Others, using calculations arranged in ways inspired by neurons in the human brain, can even perform deep learning with multiple levels of calculations. Machine learning Deep learning ○ Online chatbots use natural language processing (NLP) to analyze poorly typed or spoken questions, then predict which answers to give on topics ranging from shipping or business hours to merchandise and sizes. Vision recognition ○ AI helps doctors identify serious diseases based on unusual symptoms and early-warning signs, and it reads speed limit and stop signs as it guides cars through traffic. How do AI services calculate? And, Fraud detection what do they do with those ○ AI analyzes patterns calculations? created when thousands of bank customers make Analysis - AI services can take in (or credit card purchases, “ingest”) enormous amounts of data. then predicts which They can apply mathematical charges might be the calculations in order to analyze data, result of identity theft. sorting and organizing it in ways that would have been considered Today’s AI has gone beyond creating impossible only a few years ago. driving directions, vacuuming floors, or recommending new fashions. Now it Prediction - AI services can use their really can mimic the capabilities of the data analysis to make predictions. They human mind. AI can learn from can, in effect, say, “Based on this examples and experience, recognize information, a certain thing will objects, understand and respond to probably happen.” language, and solve problems. Even more exciting are its future What predictions can AI possibilities. make? How is AI evolving? Autocorrect spelling Human Language Computer scientists have identified three levels of AI based on predicted growth in its ability to analyze data and more versatile and can handle a make predictions. They call these wider range of related tasks. levels: Broad AI is focused on integrating AI within a specific Narrow AI business process where Broad AI companies need business- and General AI enterprise-specific knowledge and data to train this type of system. Newer Broad AI systems predict global weather, trace pandemics, and help businesses predict future trends. General AI Narrow AI General AI refers to machines Narrow AI is focused on that can perform any addressing a single task such as intellectual task that a human predicting your next purchase can. or planning your day. Currently, AI does not have the Narrow AI is scaling very quickly ability to think abstractly, in the consumer world, in which strategize, and use previous there are a lot of common tasks experiences to come up with and data to train AI systems. For new, creative ideas as humans example, you can buy a book do, such as inventing a new with a voice-based device. product or responding to people Narrow AI also enables robust with appropriate emotions. And applications, such as using Siri don't worry, AI is nowhere near on an iPhone, the Amazon this point. recommendation engine, autonomous vehicles, and more. There might be another level, known Narrow AI systems like Siri have as artificial superintelligence (ASI) conversational capabilities, but that could appear near the end of this only if you stick to the script. century. Then machines might become self-aware! Even then, no Broad AI levels of AI are expected to replace or Broad AI is a midpoint between dominate you. Narrow and General AI. Rather than being limited to a single task, Broad AI systems are Module 2: What are the found patterns and structure within the data—useful three eras of computing? meaning beyond mere The Era of Tabulation numbers. These machines uncovered ways that different Dark data - It’s information without a groups within the population structure: just a huge, unsorted mess moved and settled, earned a of facts. living, or experienced health problems—information that To sort out unstructured data, helped governments better humans have created many different understand and serve them. calculating machines. The word to remember across those Over 2000 years ago, tax twenty centuries is tabulate. Think of collectors for Emperor Qin tabulation as “slicing and dicing” data Shihuang used the abacus—a to give it a structure, so that people device with beads on wires—to can uncover patterns of useful break down tax receipts and information. You tabulate when you arrange them into categories. want to get a feel for what all those From this, they could determine columns and rows of data in a table how much the Emperor should really mean. spend on building extensions to Researchers call these centuries the the Great Wall of China. Era of Tabulation, a time when In England during the machines helped humans sort data mid-1800s, Charles Babbage into structures to reveal its secrets. and Ada Lovelace designed (but never finished) what they called The Era of Programming a “difference engine” designed to handle complex calculations Data analysis changed in the 1940s using logarithms and During the turmoil of World War trigonometry. Had they built it, II, a new approach to dark data the difference engine might emerged: the Era of have helped the English Navy Programming. Scientists began build tables of ocean tides and building electronic computers, depth soundings that could like the Electronic Numerical guide English sailors through Integrator and Computer rough waters. (ENIAC) at the University of By the early 1900s, companies Pennsylvania, that could run like IBM were using machines more than one kind of to tabulate and analyze the instruction (today we call those census numbers for entire “programs”) in order to do more national populations. They than one kind of calculation. didn’t just count people. They ENIAC, for example, not only  calculated artillery firing tables for the US Army, it worked in secret to study the feasibility of thermonuclear weapons. This was a huge breakthrough. Programmable computers guided astronauts from Earth to the moon and were Turing Test - "can machines think?" ; if reprogrammed during Apollo a computer can demonstrate the same 13’s troubled mission to bring its intelligence (or the results of the same astronauts safely back to Earth. intelligence) as a human. You’ve grown up during the Era of Programming. It even drives the phone The Era of AI began one summer in you hold in your hand. But the dark 1956 data problem has also grown. Modern Early in the summer of 1956, a businesses and technology generate small group of researchers, led so much data that even the finest by John McCarthy and Marvin programmable supercomputer can't Minsky, gathered at Dartmouth analyze it before the “heat-death” of College in New Hampshire. the universe. Electronic computing is There, at one of the oldest facing a crisis. colleges in the United States, they launched a revolution in The Era of AI scientific research and coined the term “artificial intelligence”. The history of artificial intelligence The researchers proposed that dates back to philosophers thinking “every aspect of learning or any about the question, "What more can other feature of intelligence can be done with the world we live in?" be so precisely described that a This question lead to discussions and machine can be made to the very beginning of many ideas simulate it.” They called their about the possibilities involving vision “artificial intelligence” and technology. they raised millions of dollars to achieve it within 20 years. Since the advent of electronic During the next two decades, computing, there are some important they accomplished tremendous events and milestones in the evolution things, creating machines that of artificial intelligence to know about. could prove geometry theorems, speak simple English, and even solve word problems with algebra. But then came winter  By the early 1970s, it became systems. Again, funding for AI clear that the problem was collapsed. larger than researchers imagined. There were Then came another AI chill fundamental limits that no In the late 1980s, the boom in AI amount of money and effort research cooled, in part, because could solve. of the rise of personal ○ Limited calculating power computers. Machines from ○ Limited information Apple and IBM, sitting on desks storage in people’s homes, grew more For a short time, AI was one of powerful than the huge the most exciting fields in corporate systems purchased computer science. just a few years earlier. As these issues became clear, Businesses and governments the money dried up for The stopped investing in large-scale First Winter of AI. computing research, and funding dried up. The weather was rough for half a Over 300 AI companies shut century down or went bankrupt during It took about a decade for The Second Winter of AI. technology and AI theory to catch up, primarily with new Now, the forecast is sunny forms of AI called “expert In the mid-1990s, almost half a systems”. These were limited to century after the Dartmouth specific knowledge that could research project, the Second be manipulated with sets of Winter of AI began to thaw. rules. They worked well Behind the scenes, computer enough—for a while—and processing finally reached became popular in the 1980s. speeds fast enough for Money poured in. Researchers machines to solve complex invested in tremendous problems. mainframe machines that cost At the same time, the public millions of dollars and occupied began to see AI’s ability to play entire floors of large university sophisticated games. and corporate buildings. It ○ In 1997, IBM’s Deep Blue seemed as if AI was booming beat the world’s chess once again. champion by processing over 200 million possible But soon the needs of scientists, moves per second. businesses, and governments ○ In 2005, a Stanford outgrew even these new University robot drove  itself down a 131-mile or Microsoft Excel. Examples of desert trail. structured data includes names, ○ In 2011, IBM’s Watson dates, addresses, credit card defeated two grand numbers, stock information. champions in the game Unstructured data, also known of Jeopardy! as dark data, is typically categorized as qualitative data. Today, AI has proven its ability in fields It cannot be processed and ranging from cancer research and big analyzed by conventional data data analysis to defense systems and tools and methods. energy production. Artificial Unstructured data lacks any intelligence has come of age. AI has built-in organization, or become one of the hottest fields of structure. Examples of computer science. Its achievements unstructured data include impact people every day and its images, texts, customer abilities increase exponentially. The comments, medical records, Two Winters of AI have ended! and even song lyrics. Semi-structured data is the Module 3: Structured, “bridge” between structured and unstructured data. It semi-structured or doesn't have a predefined data unstructured data: What model. It combines features of are the differences? both structured data and unstructured data. It's more A look at the types of data complex than structured data, yet easier to store than Data is raw information. Data might unstructured data. be facts, statistics, opinions, or any kind Semi-structured data uses of content that is recorded in some metadata to identify specific format. This could include voices, data characteristics and scale photos, names, and even dance moves! data into records and preset fields. Metadata ultimately Data can be organized into the enables semi-structured data to following three types. be better cataloged, searched, Structured data is typically and analyzed than categorized as quantitative unstructured data. An example data and is highly organized. of semi-structured data is a Structured data is information video on a social media site. The that can be organized in rows video by itself is unstructured and columns. Perhaps you've data, but a video typically has seen structured data in a text for the internet to easily spreadsheet, like Google Sheets categorize that information,  such as through a hashtag to big city traffic using a navigation identify a location. system. It’s a dark data problem because solving it requires working Experts estimate that about 80% of all with not only a complicated street the data in today’s world is map, but also with changing variables unstructured. It contains so many like weather, traffic jams, and variables and changes so quickly that accidents. Let’s look at how two no conventional computer program different systems might try to solve can learn much from it. this problem Analyzing Unstructured Data Programmable computer But AI can shed light on unstructured data! AI uses new kinds of computing—some modeled on the human brain—to rapidly give dark data structure, and from it, make new discoveries. AI can even learn things—by itself—from the data it manages and teach itself how to make better predictions over time. This is the Era of AI, and it changes everything! Module 4: Is machine learning the answer to the unstructured data problem? How does machine learning AI with machine learning approach a problem? Two ways to solve dark data problems If AI doesn’t rely on programming instructions to work with unstructured data, how does AI do it? Machine learning can analyze dark data far more quickly than a programmable computer can. To see why, consider the problem of finding a route through  on traffic right now, this route is likely to be faster than that one.” It knows this because it compared routes as it built them. Machine learning learns! It can notice that your car was delayed by a temporary detour and adjust its recommendations to help other drivers. The machine learning process is Machine learning uses entirely different probabilistic calculation There are two other ways to contrast The machine learning process has classical and machine learning advantages: systems. One is deterministic and the It doesn’t need a database of all other is probabilistic. the possible routes from one place to another. It just needs to Deterministic know where places are on the there must be an enormous, map. predetermined structure of It can respond to traffic routes—a gigantic database of problems quickly because it possibilities from which the doesn’t need to store alternative machine can make its choice. If routes for every possible traffic a certain route leads to the situation. It notes where destination, then the machine slowdowns are and finds a way flags it as “YES”. If not, it flags it around them through trial and as “NO”. This is basically binary error. thinking: on or off, yes or no. This It can work very quickly. While is the essence of a computer trying single turns one at a time, program. The answer is either it can work through millions of true or false, not a confidence tiny calculations. value. Probabilistic But machine learning has two more It never says “YES” or “NO” advantages that programmable Machine learning is analog computers lack: (like waves gradually going up and down) rather than binary Machine learning can predict. (like arrows pointing upward You know this already. A and downward). machine can determine, “Based  Machine learning constructs Machine learning enables a rich every possible route to a partnership between technology and destination and compares them humans in real time, including all the variables such as changing AI systems and humans excel at traffic. So, a machine learning different things. For example, you, as a system doesn’t say, “This is the person, might excel at imagining fastest route.” It says something possibilities, while AI excels at like, “I am 84% confident that pinpointing patterns. this route will get you there in the shortest time.” You might Does common sense make sense? have seen this yourself if you’ve It turns out that in fields, ranging from traveled in a car with an medicine and education to social up-to-date GPS navigation studies and government, the best system that offers you two or decisions are made using a balance of three choices with estimated human and machine strengths. But times. remember, there is another elusive but vital capability that must also be If machine learning offers only considered: common sense. You probabilities, who makes the final might know people with strong decision? common sense and understand its value. You also might have seen or This can literally be a life-and-death read output from machines that question. Suppose you have a serious makes no sense. Yet, there’s a disease and your doctor offers you a contribution to be made from both choice. Do you want your doctor to sides. prescribe your treatment, or do you want the treatment that a machine Common sense draws on many learning system determines is most complex generalizations mixed with likely to succeed? compassion and abstractions. At this time, only humans can use common sense well. The problem is that common sense is often tainted with bias that can distort your judgment. AI systems can balance this. As long as AI systems are provided and trained with unbiased data, they can make recommendations that are free of bias. A partnership between humans and machines can lead to sensible decisions. Module 5: Three common Example methods of machine The machine might be fed many learning photos and articles about dogs. It will classify and cluster information about Supervised learning all of them. When shown a new photo Supervised learning of a dog, the machine can identify the providing AI with enough photo as a dog, with reasonable examples to make accurate accuracy. predictions. Unsupervised learning occurs when All supervised learning algorithms the algorithm is not given a specific need labeled data. Labeled data is “wrong” or “right” outcome. Instead, data that is grouped into samples that the algorithm is given unlabeled data. are tagged with one or more labels. In other words, applying supervised Unsupervised learning is helpful when learning requires you to tell your you don't know how to classify data. model: Example: imagine you work for a banking institution and you have a What the key characteristics of large set of customer financial data. a thing are, also called features You don't know what type of groups or What the thing actually is categories to organize the data. Here, an unsupervised learning algorithm Example could find natural groupings of similar The information might be drawings customers in a database, and then you and photos of animals, some of which could describe and label them. are dogs and are labeled “dog”. The machine will learn by identifying a This type of learning has the ability to pattern for “dog”. When the machine discover similarities and differences in sees a new dog photo and is asked, information, which makes it an ideal “What is this?”, it will respond, “dog”, solution for exploratory data analysis, with high accuracy. This is known as a cross-selling strategies, customer classification problem. segmentation, and image recognition. Unsupervised learning Reinforcement learning Unsupervised learning Reinforcement learning a person feeds a machine a a machine learning model large amount of information, similar to supervised learning, asks a question, and then the but the algorithm isn’t trained machine is left to figure out how using sample data. This model to answer the question by itself. learns as it goes by using trial  and error. A sequence of Module 6: How will successful outcomes is reinforced to develop the best machine learning recommendation for a given transform human life? problem. The foundation of reinforcement learning is Take another look at the three rewarding the “right” behavior levels of artificial intelligence and punishing the “wrong” Think about this behavior. Perhaps, 25 years from now, General AI is expected to Rewarding a machine means emerge. AI researcher Nick that you give your agent positive Bostrom defines this reinforcement for performing superintelligence as, “an the "right" thing and negative intellect that is much smarter reinforcement for performing than the best human brains in the "wrong" things. practically every field, including scientific creativity, general As a machine learns through trial and wisdom and social skills.” error, it tries a prediction, then You’re likely to see General AI compares it with data in its corpus. appear in your lifetime. General AI will enable supersmart bots Each time the comparison is and technologies to link AI with positive, the machine receives the Internet of Things through positive numerical feedback, or “embodied cognition”. This will a reward. give machines the ability to Each time the comparison is interact in human-like ways as negative, the machine receives they work alongside humans. negative numerical feedback, or a penalty. What will interacting with general AI feel like to humans? Over time, a machine’s predictions will AI everywhere grow to be more accurate. It AI will move into all industries, from accomplishes this automatically finance, to education, to healthcare. AI based on feedback, rather than will increase productivity and enable through human intervention. new opportunities. Deeper insights New technologies will sense, analyze, and understand things never before possible. Engagement reimagined New forms of human-machine 3) The three levels of AI include: interaction and emerging Narrow AI, Broad AI, and General technologies, such as conversational AI. Narrow AI and Broad AI are bots, will transform how humans available today. In fact, most engage with each other and with enterprises use Broad AI. machines. General AI won’t come online Personalization until sometime in the future. Machine interactions will be 4) The history of AI has progressed personalized for you, with new levels of across the Era of Tabulation, detail and scale. Era of Programming, and Era of Instrumented planet AI. Billions of sensors generating exabytes 5) Data can be structured, of data will open new possibilities for unstructured, or improving Earth’s safety, sustainability, semi-structured. and security. a) Structured data is quantitative and highly What’s beyond these wonders? organized, such as a Humans, devices, and robots might spreadsheet of data. exist as a collective “digital brain” that b) Unstructured data is anticipates human needs, makes qualitative data that predictions, and provides solutions. doesn't have structure, Farther in the future, we might trust such as medical records. the digital brain to do things on our It's becoming increasing behalf across a broad spectrum of valuable to businesses. endeavors! c) And semi-structured data combines features Key points to remember of both structured data and unstructured data. It 1) Artificial intelligence refers to uses metadata. the ability of a machine to learn 6) About 80% of all the data in patterns and make predictions. today’s world is unstructured. AI does not replace human 7) Machine learning has decisions; instead, AI adds value advantages compared to to human judgment. programmable computers. 2) AI performs tasks without Machine learning can predict human intervention and and machine learning learns! completes mundane and 8) Machine learning uses three repetitive tasks, while methods. augmented intelligence allows a) Supervised learning humans to make final decisions requires enough after analyzing data, reports, examples to make and other types of data. accurate predictions  b) Unsupervised learning requires large amounts of information so the machine can ask a question, and then figure out how to answer the question by itself. c) Reinforcement learning requires the process of trial and error. 9) With AI everywhere, AI will move into all industries, from finance, to education, to healthcare. 10) AI can increase productivity, create new opportunities, provide deeper insights, and enable personalization.