Mini-lecture on Wisdom of Crowds and Collective Intelligence
Collective Intelligence
Idea that groups can be smarter than the smartest person in the group
Sometimes it’s simple; e.g, aggregating information from multiple sources
Sometimes there is an emergent property, where the group thinks and behaves differently
Ants
Organizations
While organizations are much more hierarchical than ants, they can still have emergent properties
Partially a function of history, partially of the structure of the organization
Even a culture everyone dislikes can be hard to change
E.g., metric system, QWERTY keyboards
Cities
High-level decisions about zoning, infrastructure, etc. but the character of a city is emergent
Jane Jacobs: “The Death and Life of Great American Cities”
Wisdom of Crowds Activity
Get out a piece of paper
Guess how many pages are in this book and write down your guess
Tell your guess (and no other information) to the two people on either side of you
Guess again how many pages are in the book and write it on a new line (Guess 2)
Activity 2
Guess 3: Guess the pages in this book
Two random people will tell us their guess
Guess 4: Guess the pages again for this book
Collective understanding and decision-making
Wisdom of crowds is idea that in many contexts, the average is a really good guess!
What is groupthink? Have you seen examples? How might this relate?
Dr. Josh Becker et al. show that in some cases, communication can help
Those whose guesses are worst move the most
Divorced from other social relationships (not multiplex)
Results
Decentralized network
Centralized network
Connections to other network concepts
How does this relate to our week on small group networks?
Transactive memory
Hierarchy
Centralization
Connections to other network concepts
How could you design a group / network to get the benefits of the wisdom of crowds?
When would you expect a wisdom of crowds approach to work well?
When would you expect it to fail?
Discussion Questions
What are the most practical ways to design communication networks or platforms that promote decentralized influence?
What are some real-life examples of collective intelligence making smart decisions regarding norms, such as when IP implicitly became the norm for network communication, as mentioned in the discussion video?
If collective intelligence relies on social structures more than on individual skills, then what might schools, organizations, or communities do to foster environments that promote informed participation and the real “wisdom of the crowd”?
Plan for Today
Exam Review
Exam Review
Visualization / ggraph
Look at code and explain what it will look like
Find the bug in code
Example
What code produces this graph?
G |>mutate(centrality =centrality_degree()) |>activate(edges) |>ggraph() +geom_edge_fan(color ='purple') +geom_node_point(aes(size = centrality), color ='green')
Exam Review
What would this code do?
G |>activate(nodes) |>mutate(ec =centrality_eigen()) |>filter(ec >1) |>ggraph() +geom_edge_fan(color ='orange') +geom_node_point(aes(size = ec), color ='green')
Exam Review
Centrality
Network representations (Edgelists and matrices)
Network formation
Homophily, focus theory, transitivity, social exchange theory, etc.