Chapter 11: Managing Knowledge and Artificial Intelligence Flashcards
role of knowledge management systems, AI & ML, enterprise, knowledge work systems
Managing AI: Grand vision vs Realistic vision
- grand vision: computer hardware & software that are as “smart” as humans, so far this is not the case - at least holistically speaking
- realistic vision: systems take data inputs, process them & produce outputs, but on a complexity level that is far superior to human capabilities
Major types of AI
- expert systems
- machine learning
- neural networks & deep learning networks
- genetic algorithms
- natural language processing (LLMs)
- computer vision
- robotics
Expert systems
- capturing tacit knowledge in very specific & limited domain of human expertise, knowledge as set of rules
- performing limited tasks- discrete, highly structured DM
- knowledge base: set of hundreds or thousands of rules
- inference engine: Strategy used to search knowledge base, using forward chaining (data-driven technique towards a goal) or backward chaining (goal-driven technique from a goal towards origin to extract facts)
Machine learning
- Software that can identify patterns & relationships in very large data sets without programming, but significant human training
- pattern recognition
- requires experience through prior learnings (database)
- supervised learning turns into unsupervised learning
–> google searches, recommender systems
–> sustainability issue
Neural networks
- definition: algorithms loosely based on the processing patterns of the biological brain that can be trained to classify objects into known categories based on data inputs
- purpose: inding patterns and relationships in massive amounts of data
- functioning: “learn” patterns by searching for relationships, building models & correcting repeatedly
- data input: humans “train” network by feeding it data inputs for which outputs are known
- input layer –> hidden layer –> output layer
Genetic algorithms
- problem-solving methods that promote the evolution of solutions to specified problems using the model of living organism adapting to their environment (evolution)
- useful for finding optimal solution for specific problem by examining very large number of possible solutions (used in optimization problems)
Natural language processing
- understand & speak in natural language, read natural language & translate
- typically today based on machine learning, aided by very large databases of common phrases & sentences in a given language
- Google Translate, spam filtering systems, customer call center interactions, assistances (Siri, Alexa, Cortana, Google Assistant)
Computer vision systems
- digital image systems that create a digital map of an image (like a face, or a street sign)
- Facebook’s DeepFace, passport control, identifying people in crowds, autonomous vehicles, industrial machine (robot) vision
Robotics
- design, construction, and operation of machines that can substitute for humans in many factory, office & home applications (home vacuums)
- generally programmed to perform specific and detailed actions in limited domains
- Used in dangerous situations like bomb disposal
- Surgical robots are expanding their capabilities
Intelligent agents
- work without direct human intervention to carry out repetitive, predictable tasks (limited built-in or learned knowledge base) –> chatbots, agent-based modeling applications
What is hybrid intelligence & what is it used for? (NOT RELEVANT)
„ability to achieve complex goals by combining human and artificial intelligence, thereby reaching superior results to those each of them could have accomplished separately, and continuously improve by learning from each other“
- collectivity, superior results, continuous learning
- used when the result of joint work is better than the separate work
- most probable paradigm for division of human & machine labor, as it highlights complementary strenghts (contrary to collective intelligence)
Different types of intelligence (NOT RELEVANT)
- intelligence: „the ability to accomplish complex goals, learn, reason, and adaptively perform effective actions within an environment” –> acquire and apply knowledge
- collective intelligence: „groups of individuals acting collectively in ways that seem intelligent’‘
- artificial intelligence: ‘‘[…] systems that perform activities that we associate with human thinking, activities such as decision-making, problem solving, learning“
What is the Moravec paradox? (NOT RELEVANT)
It is comparatively easy to make coputers exhibit adult level performance on intelligence tests etc., but the skills of a 1-year old (perception, mobility, common sense) are difficult to emulate
What is the “human in the loop”? (NOT RELEVANT)
integration of a human workforce in the AI pipeline in order to train and validate models in a continuous way
Relevance of knowledge management systems in business
- fastest growing areas of software investment
- 37% U.S. labor force: knowledge and information workers
- 55% U.S. GDP from knowledge and information sectors
- Substantial part of a firm’s stock market value is related to intangible assets: knowledge, brands, reputations, and unique business processes
Dimensions of knowledge
- (raw) data, knowledge & wisdom (where, when & how to apply knowledge)
- tacit knowledge: the knowledge that is implicit in a person’s actions without them being able to describe their skills verbally
- explicit knowledge: knowledge that can be readily articulated, conceptualized, codified, formalized, stored and accessed
- knowledge is a firm asset, has different forms, has a location & is situational
- Knowing how to do things effectively and efficiently in ways others cannot duplicate is a prime source of profit and competitive advantage (example: Having a unique build-to-order production system)
Define knowledge management
Set of business processes developed in an organization to create, store, transfer & apply knowledge
Define organizational learning
Process in which organizations gain experience through collection of data, measurement, trial and error & feedback
What is the knowledge management value chain?
Knowledge management today involves both information systems activities and a host of enabling management and organizational activities. 4 stages:
1. knowledge acquisition
- documenting tacit & explicit knowledge, creating knowledge, tracking data from TPS & external sources
2. knowledge storage
- databases, document management systems, role of management
3. knowledge dissemination
- portals, wikis, e-mail, instant messaging, search engines, collaboration tools, tools to focus attention on important information
4. knowledge application
- new business practices, new products & services, new markets
What measures can beimplemented to build org & management capital?
- Developing new organizational roles and responsibilities for the acquisition of knowledge
- Chief knowledge officer executives
- Dedicated staff / knowledge managers
- Communities of practice (COPs): Informal social networks of professionals and employees, Activities including education, online newsletters, sharing knowledge, reducing learning curves of new employees
Types of knowledge management systems
- Enterprise-wide knowledge management systems: General-purpose firm-wide efforts to collect, store, distribute, and apply digital content and knowledge
- Knowledge work systems (KWS): Specialized systems built for engineers, scientists, other knowledge workers charged with discovering & creating new knowledge
- Intelligent techniques: Diverse group of techniques such as data mining used for various goals: discovering knowledge, distilling knowledge, discovering optimal solutions
Types of knowledge in an enterprise
- structured documents: reports, presentations, formal rules (only 20% in a business)
- semistructured documents: e-mails, videos
- unstructured, tacit knowledge
Use cases of AI
- Recognize millions of faces in seconds
- Interpret millions of CT scans in minutes
- Analyze millions of financial records
- Detect patterns in very large Big Data databases
- Improve their performance over time (“learn”)
- Navigate a car in certain limited conditions
- Respond to questions from humans (natural language), speech activated assistants like Siri, Alexa, and Cortana
What does an enterprise content management system do?
- capabilities for classifying, organizing & managing structured and semistructured knowledge and making it available throughout the enterprise
- bring in external resources, such as news feeds & research
- tools for communication & collaboration
- key problem: developing taxonomy
- solution: digital asset management (DAM) systems
What is a DAM system?
- used by companies to store, organize, locate and share digital files.
- granting employees as well as internal & external stakeholders access to the content of a brand’s digital media library, i.e. its images, videos, presentations, documents and other digital assets
Locating & sharing expertise through knowledge management systems
- Provide online directory of corporate experts in well-defined knowledge domains
- Search tools enable employees to find appropriate expert in a company –> Problem: access rights
- Social networking and social business tools for finding knowledge outside the firm (Saving, Tagging, Sharing web pages)
What does a learning management system (LMS) do?
- Provide tools for management, delivery, tracking & assessment of employee learning and training
- Support multiple modes of learning (CD-ROM, web-based classes, online forums etc.)
- Automates selection & administration of courses
- Assembles & delivers learning content
- Measures learning effectiveness
- Massively open online courses (MOOCs) (many participants)
- Goal for the future: Personalized LMS or even individualized learning journey
What are knowledge workers & what are their roles?
Researchers, designers, architects, scientists, engineers who create knowledge for the
organization, 3 key roles:
1. Keeping organization current in knowledge
2. Serving as internal consultants regarding their areas of expertise
3. Acting as change agents, evaluating, initiating, and promoting change projects
Define knowledge work systems
Specialized systems for knowledge workers to help create new knowledge and integrate that knowledge into
business
Name the requirements of knowledge work systems (6)
- Sufficient computing power for graphics, complex calculations
- Powerful graphics and analytical tools
- Communications & document management
- Access to external databases
- User-friendly interfaces
- Optimized for tasks to be performed (design engineering, financial analysis)
Examples of knowledge work systems
- CAD (computer-aided design): Creation of engineering or architectural designs, 3D printing
- Virtual reality systems: Simulate real-life environments, 3D medical modeling for surgeons, Augmented reality (AR) systems, VRML
Name intelligent techniques to capture individual & collective knowledge to extend knowledge base
- capturing tacit knowledge: Expert systems, case-based reasoning, fuzzy logic
- Knowledge discovery: Neural networks & data mining
- Generating solutions to complex problems: Genetic algorithms
- Automating tasks: Intelligent agents
