Wikipedia:Wiki Ed/Columbia University/Computational Statistics (Spring 2016)

From Wikipedia, the free encyclopedia
This Course Wikipedia Resources Connect
Questions? Ask us:

contact@wikiedu.org

Course name
Computational Statistics
Institution
Columbia University
Instructor
Jose Blanchet
Wikipedia Expert
Ian (Wiki Ed)
Subject
Computational Statistics
Course dates
2016-01-19 – 2016-05-20
Approximate number of student editors
20


This course covers a range of tools and techniques that are used in modern statistical computations. We will start with simple randomized algorithms, in particular, basic Monte Carlo methods and output analysis. This will be covered in the first part of the course. The second part of the course involves large scale deterministic computational procedures. Finally, we will discuss hybrid methods.

Student Assigned Reviewing
LiangZhang321
Yp2342
Robinwinstanley
Amoretti86
HalJCooper
Lx2170
Zm2170
David.fei.he
YTSafire
Yh2687
Sbronder
Elbr30
Mjb2248
Weichi Yao
Hl2902

Timeline

Week 1

Course meetings
Monday, 25 January 2016   |   Wednesday, 27 January 2016
Assignment - Simulation by Inversion and Poisson Thinning

Two students selected from class will write (each) lecture notes on the topics in the title of the assignment. I'll provide editorial comments. These notes will be used to used to comment on and (if needed) enrich existing Wikipedia entries.

Milestones

The students have been given a list of topics that will be covered in the course. During this week, the students should Google the topics that we will cover. These include: 
Basic simulation techniques (such as inversion, thinning, and acceptance / rejection); variance reduction techniques; Markov chain Monte Carlo (including Metropolis-Hastings, Gibbs sampler, annealing and tempering); Rates of convergence to stationarity and simulation output analysis; Particle methods (for filtering) - also known as sequential importance sampling; numerical integration methods (introduction to Quasi-Monte Carlo); Optimization (LP and convex); EM algorithm and applications to Maximum Likelihood estimation; numerical optimization (gradient descent, conjugate gradient, and Newton steps); stochastic approximation techniques. 

The students should explore what is out there in Wikipedia with regard to these topics.

Week 2

Course meetings
Monday, 1 February 2016   |   Wednesday, 3 February 2016
Assignment - Acceptance Rejection (Rejection Sampling) and Applications

Two students selected from class will write (each) lecture notes on the topics in the title of the assignment. I'll provide editorial comments. These notes will be used to used to comment on and (if needed) enrich existing Wikipedia entries.

Milestones

Students to continue exploring Wikipedia with regard to topics of the class. At this point they should be forming an idea of what do they want to explore in order to write an entry.

Play with your Sandbox in your Wikipedia page.  

Week 3

Course meetings
Monday, 8 February 2016   |   Wednesday, 10 February 2016
Assignment - Variance Reduction Techniques I
Control Variates & Conditioning

Two students selected from class will write (each) lecture notes on the topics in the title of the assignment. I'll provide editorial comments. These notes will be used to used to comment on and (if needed) enrich existing Wikipedia entries.

Milestones

Students must pick a topic at this point. They will start collecting relevant references based on the textbooks and additional material given in class (Columbia\Courseworks), and feedback given in office hours.

Week 4

Course meetings
Monday, 15 February 2016   |   Wednesday, 17 February 2016
Assignment - Importance Sampling and Applications

Two students selected from class will write (each) lecture notes on the topics in the title of the assignment. I'll provide editorial comments. These notes will be used to used to comment on and (if needed) enrich existing Wikipedia entries.

Week 5

Course meetings
Monday, 22 February 2016   |   Wednesday, 24 February 2016
In class - Start with the structure of your article
Ask relevant questions

The goal is to have an idea of what's the ideal content (to a first approximation) of the article.

What is the motivation of the method?
Maybe think of a simple example that will guide your understanding of how the method is applied. At this point you may NOT know how the method actually works because this will come later in the course. Ultimately here you'll want to answer: How does the method work?
Why the method works? Again, you might not answer this, but you can start searching for references that you might find "easy" to understand. 

Collect references for these basic questions. Ask questions in class!

Week 6

Course meetings
Monday, 29 February 2016   |   Wednesday, 2 March 2016
Assignment - Bias Reduction (Deletion) Techniques
Multilevel Monte Carlo

Two students selected from class will write (each) lecture notes on the topics in the title of the assignment. I'll provide editorial comments. These notes will be used to used to comment on and (if needed) enrich existing Wikipedia entries.

Milestones

You should be able to start writing some of the content of the article. Specially, you must play with the mathematical equations and figures. Use the examples that you identified in the previous weeks. Discuss the examples, again, you might not know the methods at this point, but you can write the motivation, insert references, describe the example mathematically, provide a figure for a plot or something like that.

The first part of this course deals with Monte Carlo, and the second part mostly with Optimization. By this time you'll already have enough elements to actually fully explain "easy" examples. You can do so, not necessarily using the method that you have selected, but to practice editing articles.

Week 7

Course meetings
Monday, 7 March 2016   |   Wednesday, 9 March 2016
Assignment - Markov chain Monte Carlo 1
Metropolis Hastings

Two students selected from class will write (each) lecture notes on the topics in the title of the assignment. I'll provide editorial comments. These notes will be used to used to comment on and (if needed) enrich existing Wikipedia entries.

In class - Editing previous entries, reading, and working on article

This week and the next one are crucial. If you have fallen a bit behind, this is the time to catch up. Continue practicing by editing previous Wikipedia articles that relate to ALL of the topics that have been covered in class. You should be able to help others in class, by checking out their articles and by providing useful feedback. 

The next week will be Spring break at Columbia. Read on your own, prepare questions to improve your article, try to do as much as possible now!

Hopefully, at this point you should be able to move your article to "Main Space". 

Week 8

Course meetings
Monday, 21 March 2016   |   Wednesday, 23 March 2016
Assignment - Markov chain Monte Carlo
Gibbs and Other Methods

Two students selected from class will write (each) lecture notes on the topics in the title of the assignment. I'll provide editorial comments. These notes will be used to used to comment on and (if needed) enrich existing Wikipedia entries.

In class - Give feedback and provide feedback

By this time, everyone must have an article in Mainspace. So every one should be able to get feedback, provide feedback, and respond to it.

Week 9

Course meetings
Monday, 28 March 2016   |   Wednesday, 30 March 2016
Assignment - Rates of Convergence to Stationarity and Output Analysis

Two students selected from class will write (each) lecture notes on the topics in the title of the assignment. I'll provide editorial comments. These notes will be used to used to comment on and (if needed) enrich existing Wikipedia entries.

Milestones

Most people at this time will have the main elements to write their article. So, from this time on you should work on improving it, adding examples, explaining things better, adding figures and diagrams, references, etc.

Week 10

Course meetings
Monday, 4 April 2016   |   Wednesday, 6 April 2016
Assignment - Sequential Importance Sampling and Particle Methods

Two students selected from class will write (each) lecture notes on the topics in the title of the assignment. I'll provide editorial comments. These notes will be used to used to comment on and (if needed) enrich existing Wikipedia entries.

Milestones

Everyone should now go and look at each other's articles. Provide feedback and constructive comments. If you have questions, you should bring them up either in class or office hours.

Week 11

Course meetings
Monday, 11 April 2016   |   Wednesday, 13 April 2016
In class - Optimization
Linear Programming

Two students selected from class will write (each) lecture notes on the topics in the title of the assignment. I'll provide editorial comments. These notes will be used to used to comment on and (if needed) enrich existing Wikipedia entries.

Milestones

At this point most of you will be in their final stages of your articles. Some, those who picked topics that are covered in the later parts of the course, might still be working on completing the articles. 

Continue providing feedback and improving your articles. 

Week 12

Course meetings
Wednesday, 20 April 2016   |   Monday, 18 April 2016
In class - Convex Optimization, Conjugate Gradients, etc.

Two students selected from class will write (each) lecture notes on the topics in the title of the assignment. I'll provide editorial comments. These notes will be used to used to comment on and (if needed) enrich existing Wikipedia entries.

In class - Article completion

Virtually everyone at this point should be done, or be close to done with their articles. Implement the latest rounds of feedback.

Week 13

Course meetings
Monday, 25 April 2016   |   Wednesday, 27 April 2016
Assignment - EM Algorithm and Other methods

Two students selected from class will write (each) lecture notes on the topics in the title of the assignment. I'll provide editorial comments. These notes will be used to used to comment on and (if needed) enrich existing Wikipedia entries.

In class - Last chance to catch up!

Take this week as a "buffer" to really get things done!

Week 14

Course meetings
Monday, 2 May 2016   |   Wednesday, 4 May 2016
Assignment - Stochastic Approximations

Two students selected from class will write (each) lecture notes on the topics in the title of the assignment. I'll provide editorial comments. These notes will be used to used to comment on and (if needed) enrich existing Wikipedia entries.