CMN 150V Midterm (UCD) Complete
Exam Study Questions with Correct
Answers | Latest Edition
1. What does she do? What does she compare her data to in order to identify
fraud? - ANSWER She compares it to theoretical models of "how data
would look like" if ballot box stuffing and/or vote stealing took place
2. Wait: why do we know what happened in the training data? - ANSWER
Because she created the training data herself (synthetic data), based on the
theory of what fraud looks like
3. If we find a polling location with very high voter turnout, and a very high
share of the incumbent in the vote share, given the identified decision-rule
(the lines that distinguish among different cases), how would this machine
learning algorithm categorize this new case? - ANSWER red: ballot box
stuffing
4. Here, the same decision rule (previously represented in a graph) is
represented in form of a "decision tree". Following this algorithm, if we find
a location where the incumbent has 50 % share of the votes and the turnout
is 80%, what would the machine intelligence detect? - ANSWER BBS
(Ballot Box Stuffing)
5. A "false " (-----) is an error in data reporting in which a test result
improperly indicates presence of a condition, such as fraud (the fraud
detection result is positive), when in reality it is not present - ANSWER
positive
,6. What are "false positives"? - ANSWER Claiming that something is true,
even so it is not
7. What are "false negatives"? - ANSWER Claiming that something is not
true, even so it is true in reality
8. A "false " (-----) is an error in which a test result improperly indicates no
presence of a condition (the fraud detection result is negative), when in
reality it is present. - ANSWER negative
9. How did Prof. Levin create her training that served her as benchmark for
fraud? - ANSWER She took the result from a previous election, assumed
that it was all clean, and added the known theoretical pattern of fraud (multi-
modality, snowflakes, etc)
10.Using her approach to analyze the voting patterns of the 2013 Presidential
election in Venezuela, what did Prof. Levin find? - ANSWER That some
25% of the locations could have exhibited some kind of voter fraud
11.Being representative of most Computational Social Scientists, Prof. Levin
argues that if machine learning detects something, we should go with it. -
ANSWER False, she argues to use them as guideposts, helping us to identify
places where to take a closer look.
12.What is the difference between first and second part of the course? -
ANSWER 1st looks at static snapshots of society
2nd looks at changing dynamics
13.What paradigm underlines "computational science"? - ANSWER the digital
revolution
,14.In the late 1980s, how much of all technology stored info was digital and
what percentage is it now? - ANSWER less than 1% -> 99%
15.All different kinds of social science disciplines are fundamentally interested
in what? - ANSWER how society emerges from individual parts
16.What was the main approach toward science adopted by Charles Darwin? -
ANSWER he made empirical observations and from there developed ideas
17.What did Albert Einstein do in 1905 and 1915? - ANSWER he developed
theory, not based on empirical observations, but on ideas and 1st principles
18.How does this relate to the "very short history of science" of 3 consecutive
waves that we had reviewed? - ANSWER Einstein worked on a problem
with few variables(E = mc^2) and Darwin on one with average tendencies of
many observations
19.What is the so-called "digital footprint" or "digital trace"? - ANSWER the
digital evidence you leave behind with a digital interaction
20.How do you approach science through induction? - ANSWER data-
>analyses->ideas
21.Loosely speaking, what term can you use to explain what a hypothesis is? -
ANSWER a directed bet
22.Deduction: - ANSWER theory->data
, 23.The method that Albert Einstein spearheaded in the work on relativity by
going from mathematical theories over hypothesis toward observable
phenomena is called: - ANSWER deduction
24.The unconventional name of 'glass of red wine theorizing' is used here to
characterize a very common way to deduce hypothesis in the social sciences.
Which one? - ANSWER examine the logical interrelationships among some
collection of verbal statements or thoughts
25.Which approaches are here informally/jokingly called "glass of red wine
theorizing"? - ANSWER -starting from visionary ideas and then collecting
empirical data to test the ideas with hypothesis
-starting from an intellectual conversation with leading experts and then
formulating hypothesis
26.Why is social science the most complex of all sciences? - ANSWER it is on
top of a hierarchy that subsumes other science disciplines
27.David Hilbert asked if its possible to automate the decision if something is
true or false - ANSWER Entscheidungs problem
28.Kurt Gobel showed that its not possible to find a consistent system with only
the statements - ANSWER incompleteness theorem
29.How did they fill up the map quickly - ANSWER with mobile phone trace
data
30.What does it mean that the amount of technologically stored info "doubles"
every 2-3 years? - ANSWER each 2-3 years, as much is added to what we
have accumulated since the very beginning
Exam Study Questions with Correct
Answers | Latest Edition
1. What does she do? What does she compare her data to in order to identify
fraud? - ANSWER She compares it to theoretical models of "how data
would look like" if ballot box stuffing and/or vote stealing took place
2. Wait: why do we know what happened in the training data? - ANSWER
Because she created the training data herself (synthetic data), based on the
theory of what fraud looks like
3. If we find a polling location with very high voter turnout, and a very high
share of the incumbent in the vote share, given the identified decision-rule
(the lines that distinguish among different cases), how would this machine
learning algorithm categorize this new case? - ANSWER red: ballot box
stuffing
4. Here, the same decision rule (previously represented in a graph) is
represented in form of a "decision tree". Following this algorithm, if we find
a location where the incumbent has 50 % share of the votes and the turnout
is 80%, what would the machine intelligence detect? - ANSWER BBS
(Ballot Box Stuffing)
5. A "false " (-----) is an error in data reporting in which a test result
improperly indicates presence of a condition, such as fraud (the fraud
detection result is positive), when in reality it is not present - ANSWER
positive
,6. What are "false positives"? - ANSWER Claiming that something is true,
even so it is not
7. What are "false negatives"? - ANSWER Claiming that something is not
true, even so it is true in reality
8. A "false " (-----) is an error in which a test result improperly indicates no
presence of a condition (the fraud detection result is negative), when in
reality it is present. - ANSWER negative
9. How did Prof. Levin create her training that served her as benchmark for
fraud? - ANSWER She took the result from a previous election, assumed
that it was all clean, and added the known theoretical pattern of fraud (multi-
modality, snowflakes, etc)
10.Using her approach to analyze the voting patterns of the 2013 Presidential
election in Venezuela, what did Prof. Levin find? - ANSWER That some
25% of the locations could have exhibited some kind of voter fraud
11.Being representative of most Computational Social Scientists, Prof. Levin
argues that if machine learning detects something, we should go with it. -
ANSWER False, she argues to use them as guideposts, helping us to identify
places where to take a closer look.
12.What is the difference between first and second part of the course? -
ANSWER 1st looks at static snapshots of society
2nd looks at changing dynamics
13.What paradigm underlines "computational science"? - ANSWER the digital
revolution
,14.In the late 1980s, how much of all technology stored info was digital and
what percentage is it now? - ANSWER less than 1% -> 99%
15.All different kinds of social science disciplines are fundamentally interested
in what? - ANSWER how society emerges from individual parts
16.What was the main approach toward science adopted by Charles Darwin? -
ANSWER he made empirical observations and from there developed ideas
17.What did Albert Einstein do in 1905 and 1915? - ANSWER he developed
theory, not based on empirical observations, but on ideas and 1st principles
18.How does this relate to the "very short history of science" of 3 consecutive
waves that we had reviewed? - ANSWER Einstein worked on a problem
with few variables(E = mc^2) and Darwin on one with average tendencies of
many observations
19.What is the so-called "digital footprint" or "digital trace"? - ANSWER the
digital evidence you leave behind with a digital interaction
20.How do you approach science through induction? - ANSWER data-
>analyses->ideas
21.Loosely speaking, what term can you use to explain what a hypothesis is? -
ANSWER a directed bet
22.Deduction: - ANSWER theory->data
, 23.The method that Albert Einstein spearheaded in the work on relativity by
going from mathematical theories over hypothesis toward observable
phenomena is called: - ANSWER deduction
24.The unconventional name of 'glass of red wine theorizing' is used here to
characterize a very common way to deduce hypothesis in the social sciences.
Which one? - ANSWER examine the logical interrelationships among some
collection of verbal statements or thoughts
25.Which approaches are here informally/jokingly called "glass of red wine
theorizing"? - ANSWER -starting from visionary ideas and then collecting
empirical data to test the ideas with hypothesis
-starting from an intellectual conversation with leading experts and then
formulating hypothesis
26.Why is social science the most complex of all sciences? - ANSWER it is on
top of a hierarchy that subsumes other science disciplines
27.David Hilbert asked if its possible to automate the decision if something is
true or false - ANSWER Entscheidungs problem
28.Kurt Gobel showed that its not possible to find a consistent system with only
the statements - ANSWER incompleteness theorem
29.How did they fill up the map quickly - ANSWER with mobile phone trace
data
30.What does it mean that the amount of technologically stored info "doubles"
every 2-3 years? - ANSWER each 2-3 years, as much is added to what we
have accumulated since the very beginning
