I400/590 : Large-Scale Social Phenomena 
Prof. Simon DeDeo
http://santafe.edu/~simon

Spring 2014 Informatics East, Room 122 
Mondays and Wednesdays, 4:00-5:15pm

OFFICIAL COURSE SYLLABUS AT: http://bit.ly/lssp2014

** Course description

What has become of human society since its emergence ten thousand years ago?
What will come next? This course will develop the concepts and tools necessary
to understand the emergence and dynamics of large-scale social phenomena
including cultural practices and social norms, revolutions and market crashes,
and conflict in both war and peace. We will take as case-studies both modern
(21st Century), historical (17th-20th Century), ancient (3rd-7th Century BCE),
and evolutionary-time sociotechnical systems.

Particular focus on topics newly accessible to "big data" approaches; emphasis
on developing research projects and building collaboration skills, including
mid-semester hackathon.

Prerequisites: programming skills and initiative sufficient to work with
real-world data (sample question: what is the variance in the end-of-day S&P
500 for 2005). Basic understanding of probabilities (conditional, joint).
Strong interest in the topic. Course content is expected to be
compl[i/e]mentary to Fall 2013 offerings “Collective Intelligence” and “Data
Visualization”.

** Goals

To prepare students for high-level work at the research frontier.

At the end of the course students will:

1. understand essential concepts and problems in the study of group-level
behavior, including the development of social norms, the emergence of
distributed information and cognition, the selection from multiple equilibria,
the "problems of rationality".

2. be able to use tools from (a) information theory, (b) Bayesian reasoning,
and (c) theory of computation, to phrase mathematically rigorous questions
about these concepts and problems.

3. be able to download and analyze real-world data to address and answer the
questions posed in (2), using both (a) Bayesian statistics and (b) null
hierarchies.

Strong emphasis placed on building collaborative skills: including (1) the
free sharing and development of ideas within the group, (2) distributed
leadership of research teams, (3) problem-solving and meta-problem-solving,
and (4) communication of concepts, tools, and results beyond the group to the
wider community.

** Requirements

One presentation; one joint mid-term hackathon; one final paper.

The class workload is "heavy in the middle". On-schedule progress will mean a
higher workload around the middle of term, with significantly reduced work
compared to a normal class at semester end.

* Mid-Term Hackathon

In keeping with our topic of study, we replace our mid-term exam with a
collaborative hackathon. Participants will work in groups of three to solve a
single question; each group will receive a different question.

In addition to intellectual development, an explicit goal of the hackathon is
to develop your collaborative skills, and your ability to rapid prototype --
to follow a good idea on a whim.

The time-limit is 72 hours, with solutions presented in a 20 minute talk at
the subsequent class. Provisional ratings will be informed by the fixed-point
of a recursive trust algorithm, using ratings submitted by fellow students.

A good use of 72 hours is to have three meetings: the first to discuss an
approach, or set of approaches to the question, and to (self-)assign
analytical and numerical tasks. The second to critically re-direct the joint
program given shared task outcomes. The third to consolidate results, to
assess the group's findings, and to plan and rehearse the final presentation.

* Final Paper

A serious but brief research paper, of between three and five pages (excluding
figures; preprint format), presenting novel results and clearly demonstrating
grasp of the (1) conceptual, (2) mathematical, and (3) data-analytic goals of
the course. Ideally based on the participant's presentation and hackathon.

Joint authorship of final papers is allowed only by prior arrangement. Length
of paper requirements scale exponentially. Figures (but not text) may be
shared between participants (with proper credit in caption). Final grades will
involve assessment of both quality and quantity of original work.