CMN 150V Midterm (UCD) Exam questions and correct detailed answers with rationales|

The digital revolution - correct answers What forms the basis for computational social science? 80% of information - correct answers how much of the world's information was stored in analog form in the late 1980s? over 99% - correct answers Today, how much of our information is stored digitally? It doubles every 2.5 years - correct answers how quickly does the amount of stored digital information increase? 5 ZB (5x10^21) - correct answers how much stored digital information was there in 2014? when we find new ways to process information - correct answers when does evolution happen? when society as a whole becomes indispensably dependent on digital technology - correct answers when do digital and biological merge? simplicity, averages, organized complexity - correct answers what are the three kinds of problems (Weaver)? 2 or 3 variable problems - correct answers simplicity probability theory; large numbers of variables in which the behavior of a certain individual is unknown; working with distributions - correct answers averages simultaneously dealing with a large number of interrelated factora - correct answers organized complexity organized complexity problem - correct answers what type of problem is computational social science physics, chemistry, biology, social, technological - correct answers what is the order of the hierarchy of science (bottom to top) dynamics of matter - correct answers physics inorganic compositions of atoms - correct answers chemistry organic forms of matter - correct answers biology study of networks among humans - correct answers social artificial intelligence and communication - correct answers technological looking at society as a single super organism that emerges from individuals who come from different organizations and who have different motives - correct answers social emergence quantitative differences can become qualitative changes - correct answers what did Marx say about social emergence? empirical, theoretical, analytical - correct answers what are the three parts of the scientific method? big data: our digital footprint; observation and collection of data; induction - correct answers what is the empirical scientific method in the context of computational social science? data to theory - correct answers induction phenomena/system, collect data, abstraction/analysis/model, hypothesis, theory - correct answers what is the process of induction? computational models about society used to make theories about society; computer simulations; deduction - correct answers what the theoretical scientific method in the context of computational social science? theory to data - correct answers deduction theory, hypothesis or stylized fact, abstraction/analysis/model, data collection, phenomena/system - correct answers what is the process of deduction? uses logical relationships among a collection of verbal statements or thoughts to deduce a hypothesis - correct answers red wine theorizing bridge between empirical and theoretical that tells us why something happened - correct answers what is the analytical scientific method in the context of computational social science? 2/3 psychological experiments cannot be reproduced successfully and there are not enough replication studies - correct answers what is the replication crisis? FALSE: social science can handle simple and large problems, but not realistic - correct answers (T/F) social science can handle realistic problems, but not simple and large problems TRUE: it is at the top of the hierarchy - correct answers (T/F) social science is the most complex of all sciences FALSE: social science is not good at producing reliable results - correct answers (T/F) social science is good at producing reliable results there is no way to optimize a method to find the best way of doing things - correct answers incompleteness theorem can we optimize the process of knowledge creation? - correct answers what is Hilbert's main research question? empirical digital footprint that we leave behind, which allows us to socially map out the world - correct answers big data allows us to explain social phenomena in terms of social networks, which are revealed by the digital footprint - correct answers social network analysis machine learning, natural language processing - correct answers what are the two components of artificial intelligence? learns patterns and applies them to perform functiona - correct answers what does artificial intelligence do? we have so much data that we need digital technology to analyze and find patterns in an overload of information - correct answers Machine Learning uses AI to find patterns in large amounts of natural language data - correct answers natural language processing allow us to analyze data that is not from the past by simulating events to test the effects of certain changes and scenarios - correct answers computer simulations interdependency - correct answers what does nonlinearity come from? mobile phone data, satellite data, census - correct answers where does big data come from? looks at mobile phone use through survey - correct answers digital trace data convert raw data into features - correct answers feature engineering quantitative metrics that can be input into a model - correct answers features given known inputs and outputs, the machine can learn a function that maps inputs and outputs - correct answers supervised machine learning if inputs and outputs correlate; determines outputs given inputs - correct answers what does a program model tell us? take the algorithm and surveys and see if they correlate - correct answers how can we determine accuracy? allows us to make cheaper substitutions for unavailable data - correct answers band-aid statistics allows us to detect changes in quality of life which allows for better targeting of aid, new approaches to impact evaluation, and disaster response - correct answers real time monitoring digital footprint (veracity), data fusion (variety), no sampling (volume), in real time (velocity), machine learning - correct answers what are the components of big data digital footprint - correct answers veracity data fusion - correct answers variety no sampling - correct answers volume in real time - correct answers velocity gives empirical evidence through the observation of society by digital means - correct answers digital footprint data is automatically recorded, so mobile phone companies and tech companies become data providers - correct answers who are the data providers? data about data; allows you to reverse engineer a census - correct answers metadata allows us to stitch together different aspects of the digital footprint to form a more complete picture (often messy and incomplete) - correct answers data fusion state of informal isolation that can result from personalized information exposure - correct answers filter bubble metaphorical description of a situation in which existing options are reinforced by communication - correct answers echo chamber uses whole populations (not part) in conjunction with data fusion; there is a large sample size, but it can be very biased - correct answers no sampling data is collected much more quickly than with traditional methods and it is generally available almost instantaneously, so it is dynamic - correct answers produced in real time gives automated insights using artificial intelligence to make sense of big data by looking for patterns - correct answers machine learning individualized, but takes time - correct answers content based filtering uses patterns about others to make recommendations to an individual, but it requires a lot of people - correct answers collaborative filtering there is a digital divide in that some people either don't have access to technology or choose not to use it - correct answers footprint ≠ representative predictive policing uses algorithms to guide police cars - correct answers data ≠ reality AI can be discriminatory, but that discrimination is derived from the humans who contribute to the digital footprint - correct answers meaning ≠ meaningful different recommendations may seem discriminatory, even if they are based on personal interest; it is possible to teach algorithms not to discriminate, but this affects accuracy - correct answers discrimination ≠ personalization