Final - Discussion Questions
In what ways can communications companies use geospatial analysis to harness their data effectively?
They generate massive amts of data daily. Ability to analyze it quickly + with high level of granularity = better identify customer churn and help in formatting specific strategies specific to locations for increasing operational efficiency/quality of service/revenue
How do you feel that the shift from automation to its autonomy will affect robotics development?
shift to autonomy will greatly increase possible roles for robots greatly extend types of functions they perform. on other hand, humans innately don't want to feel inferior to each other or to robots. sense of pride/self-reliances
Discuss the reasons why the classical type of expert systems is disappearing.
-acquisition of knowledge from human experts = expensive, shortage of experts, need to interview multiple experts for one app -any acquired knowledge needed to be updated frequently @ high cost -rule-based foundation was freq not robust and not reliable/flexible/had too many rule exceptions. improved systems use data-driven and stats approaches to make inferences w better results rule-based user interface needed to be supplemented (by voice/images) making it cumbersome reasoning capability of rule-based tech is limited compared to newer like machine learning
Briefly describe three benefits (process gains) derived from working in groups.
-provides learning -groups better than individuals at understanding something -people take ownership/responsibility -egos are embedded in decision making -groups better at catching errors -more information avail (knowledge, access) than any one member -produce synergy during prob solving -risk propensity is balanced (manage high risk takers with conservatives)
Compare and contrast decision making under uncertainty, risk and certainty.
1. Certainty: assumed complete knowledge avail. outcome of actions known. (deterministic) 2. Uncertainty: considers several outcomes possible. doesn't know/cant estimate probability of outcomes. more difficult bc insufficient info. 3. Risk: aka probabilistic/scholastic. must consider several outcomes for each alt, each with given probability.
List and briefly describe the major types of crowdsourcing.
1. Collective intelligence (Wisdom): crowds are solving problems and providing new insights/ideas that lead to innovations 2. Crowd Creation: people creating various types of content and sharing with others (some free, some paid). content used for solving problems, advertising, knowledge. done by splitting large tasks into small segments (ex = content on wikipedia) 3. Crowd Voting: people giving opinions/ratings on ideas/products/services. also evaluating/filtering info presented. (ex = voting for American Idol) 4. Crowd Support and Funding: people contributing and supporting endeavors for social/business causes. includes donations, micro-financing.
Discuss 4 ways that IoT can drive marketing.
1. Disruptive data collection: IoT collects more data about customers from more sources. includes wearables, smart homes, etc. changes in consumer prefs 2. Real-time personalization: can provide more accurate info about specific customer buying decisions. identify and direct customers to brands 3. Environmental attribution: can monitor environments regarding ad delivery for specific places/customers/methods/campaigns 4. Complete conversation path: initiatives expand and enrich digital channel of convos between customers and vendors. provides insight on consumer purchasing paths.
List and describe 4 categories of models. Give examples of each category.
1. Few Alternatives Optimization: find best solution from small # of alts (decision tables, decision trees, analytic hierarchy process) 2. Algorithm Optimization: find best solution from big # of alts. Uses step-by-step process. (linear/other programming models, network models) 3. Analytic Formula Optimization: find best solution in ONE step. Uses a formula. (inventory models) 4. Predictive Models Optimization: predict future for given scenario (forecasting models, Markov analytics)
List 4 rational economic assumptions the linear programming allocation model is based upon.
1. Returns from diff allocations can be compared. AKA they can be measured by common unit (dollars, utility) 2. Return from any allocation is independent of other allocations 3. all data are known WITH CERTAINTY 4. Resources to be used in most economical manner 5. total return = sum of returns yielded by different activities
List 4 things sensitivity analyses are used for.
1. Revising models to eliminate too-large sensitivities 2. Adding details about sensitive variables/scenarios 3. obtaining better estimates of sensitive ext variables 4. Altering real-world system to reduce actual sensitivities 5. Accepting/using sensitive (vulnerable) real-world. leads to continuous/close monitoring of actual results
List and describe the three main 'V's that characterize Big Data.
1. VOLUME: most common trait of Big Data. many factors created exponential increase in volume. from text analysis, streaming, social media, sensors. 2. VARIETY: data comes in diff formats (traditional DBs to hierarchy/user created OLAPs, to text, email, XML, meter-collected, sensor, video, audio, stock ticker. estimates 80-85% of all orgs' data is unstructured or semistructured. 3. VELOCITY: how fast data is produced AND how fast data must be processed to meet need/demand. RFID tags, automated sensors, GPS all drive need to deal w torrents of data in NEAR real-time
Describe and differentiate between automated buildings, smart buildings, and cognitive buildings.
AUTOMATED BUILDINGS: allow the visualization of key performance indicators and are good for automated ratings. allowing identification of general issues but DONT identify energy waste well. SMART BUILDINGS: analyze energy consumers and understand consumption of rooms in central assets well, BUT only analyze primary data points COGNITIVE BUILDINGS: designed to learn behaviors. includes predictive control down to the desk level, understanding energy flow and occupancy. Considers comfort preferences of users and can collect context in decisions.
What is NoSQL as used for Big Data? Describe its major downsides.
NoSQL is new style of db that emerged (like hadoop) to process large volumes of multi-structured data. different than hadoop bc noSQL dbs are aimed mostly at serving discrete data stored among large volumes of applications. capability is lacking from relational db technologies. cant maintain needed app perf levels at Big Data scale downside of nosql dbs is they trade ACID (atomicity, consistency, isolation, durability) compliance for performance and scalability. many also lack mature mgmt and monitoring
Describe how RFID sensors work. What is an active tag?
RFID systems consist of a tag, an interrogator, one or more antennae attached to reader, and a computer program. most retail supply chains use passive RFID tags (over electromagnetic fields) only when info is requested. field created by interrogator and backscatter information. active tags have a battery to energize themselves. active tags w/ own power source do not need a reader to energize them. they can initiate data transmission on their own
Give an example of a robot being used in a situation that would be too dangerous for humans.
Space missions Deep ocean exploration Extreme terrains and weather conditions dangerous construction locations super microscopic surgeries
Can customer relationship management (CRM) systems and revenue management systems (RMS) recommend not selling a particular product to certain customers? If so, why? If not, why not?
Yes, they can recommend not selling a particular product. Yes, can avoid recommending to certain customers. Based on if client is profitable over long time for system. Rely on forecasting/predictive analytics. Predict best/worst customers and identify what and whether or not to sell to them.
What is cloud computing? What is Amazon's general approach to the cloud computing services it provides?
a style of computing in which dynamically scalable and often virtualized resources are provided over the internet. users dont need knowledge of, experience in, or control over the infrastructure amazon is huge for ecommerce and BI, CRM, supply chain, etc. built major data centers to manage their own and others operations customers dont have to make huge initial investment bc amazon already did pay as you go subscription based use what you use
A group decision support system (GDSS) is an interactive computer-based system that facilitates the solution of semi structured or unstructured problems by a group of decision makers. What are the major characteristics of GDSS?
goal to support process of group decision makers by providing automation of sub processes using IT tools specially designed infosys (not just a config of pre-existing components) designed to address one type or variety of decisions encourages generation of ideas, conflict resolution, and freedom of expression. contains built in mechanisms to discourage development of negative group behaviors
List and briefly describe the 3 classes of bots
classified by their capabilities. 1. REGULAR BOTS: conversational intelligent agents. simple/repetitive tasks like showing bank details, help buying goods online, stock transactions. 2. CHATBOTS: more capable bots. can stimulate conversations with people. 3. INTELLIGENT BOTS: have a knowledge base that is improving with experience. they can learn a customer's preferences
Discuss some of the security concerns surrounding the Pepper robot.
concerns pointed out by scandinavian researchers easy to have unauthenticated root level access to bot prone to brute force attacks can be programmed with widely known and easily used langauges via APIs access to all sensors = not secure (mic, speakers, camera) = spying possible ongoing issues for robots/smart speakers
Describe what the IoT is.
connected network in which: large # of objects/things can be connected each thing has unique definition (IP address) each thing has ability to receive/send/store data automatically each things delivered mostly over wireless internet each built upon machine-to-machine communication
Discuss the challenges that must be considered when creating human/robot collaborations.
designing a human-machine team that capitalizes on strengths of each partner exchanging info bw humans and robots preparing employees for collaboration across the board changing processes to accommodate the collabs ensure safety of robots and humans
List at least 5 representative examples of chatbot tasks inside an enterprise.
help w/ project mgmt handle data entry conduct scheduling streamline payments advise on auth of funds monitor work/workers analyze big data find discounts simplify interactions use machine learning
Describe data stream mining and how it is used.
enabling technology for stream analytics process of extracting novel patterns and structures from continuous/rapid data records continuous flow of instances that can typically only be read once in short time period examples: sensor data, pc network traffic, phone convos, ATM transaction, web searches, financial data. stream mining = subfield of data mining/machine learning/knowledge discovery goal is to predict class/value of new instances in stream given some knowledge about the class or membership values of previous instances in stream
A group support system (GSS) is any combination of hardware and software that enhances group work either in direct or indirect support of decision making. When and how did GSS evolve?
evolved after IT researchers recognized that tech could be developed to support many activities normally happening at face-to-face meetings idea generation, consensus building, anonymous ranking
Describe why autonomous trucks would have a massive disruptive effect on jobs in the transportation industry.
it provides a huge # of jobs across USA. if autonomous were approved/deployed, industry massively impacted arguably its less expensive and safer and thus many would adopt it displaces large portion of workforce and could have negative economic effects
List and briefly describe the major components of expert systems.
knowledge acquisition: mostly from human experts. obtained by knowledge engineers. may come from several sources. integrated/validated/verified knowledge base: a repository. divided into knowledge about domain and about problem solving and procedures . input data is provided by the users knowledge representation: frequently organized as business rules (AKA production rules) Inference Engine: aka control structure or rule interpreter. this is the "brain". provides reasoning capabilities - ability to answer user's questions. engine manipulates rules by fwd/bkwd channing. 1990s User Interface: component allows user inference engine interactions. classically done by writing or using menus. now done by natural language and voice.
Paro is a social robot for older adults. What have the results of studies on this type of robot found?
results = social robots increase social interaction brings smiles to patient faces creates happy experiences for humans even tho it didnt provide complete responses that humans do, many found Paro to be meaningful and emotionally connective help break monotonous routines + add joy to older lives
List and briefly discuss some of the disadvantages and limitations of bots.
some provide inferior performance = customers frustrated initially some dont properly represent their brand. poor design = poor representation quality of AI based bots depends on use of complex algorithms which are expensive some not convenient to use some operate inconsistently security and integration challenges
List five of the major drivers of IoT
the # of things = 20 to 50 billion by mid 2020s connected autonomous things crate new IoT applications broadband internet more widely avail and always increasing cost of devices/sensors decreasing cost of connecting devices decreasing addl devices created and interconnected easily more sensors built into more devices smartphones availability of wearables convenience protocols/standards starting to develop powerful data analytics avail
