Friday, May 3, 2013

Coursera Review part 0

In this series of posts, I will share some experiences from the on-line learning platform coursera.org

1. What is coursera?

A on-line learning platform of a company founded by computer science professors Andrew Ng and Daphne Koller from Stanford University. It offers on-line university-level courses in engineering, humanities, medicine, biology, social sciences, mathematics, business, computer science, and other areas.

The courses include video lecturing, discussion groups, test quizzes, programming assignments and sometimes peer-reviewed test cases. The courses are delivered professors from by leading universities. For students successfully passing the assignments, an informal certificate is issued by coursera.

Similar platforms include: EdXStanford Venture Labs, Udacity, Khan Academy, and MIT Open Courseware.

2. Who should use it?

2.1. University students who wish to enhance their education, in particular in cases in which the student does not have the opportunity to get comparable high-quality course at his home institution.

2.2. Working professionals seeking a more mentally beneficial way to relax (in other words, an alternative to sweating in a gym or some other sport, which is also my personal motivation).

2.3. People looking for a potential career change, but uncertain about the direction which they would like to take (back-to-college attitude).

3. Who should NOT use it?

3.1. Self-learners should be careful not to place their entire plans on Coursera. Coursera cannot replace a college / university education.

3.2. People looking for purely career-enhancing options should avoid coursera at all. First, coursera offers academic courses, meaning that in the usual case only a very small fraction of what you learn will ever come in a real application. It's NOT a place for hand-on, vocational-type training! Second, the certificates issued by the platform are really silly, including all of the paid alternatives around it.

3.3. People who have difficulties in learning independently, without high degree of personal interactions.

4. What is the best thing about coursera?

It's free, high quality education!

5. What is the worst thing about coursera?

It is difficult to compose an equivalent of a bachelor degree in, say, computer science, by taking the courses offered solely on the platform (at the very least, people who are self-learning should try multiple platforms). At a graduate level, the courses offered are basically 'self-contained' for 6 to 10 weeks, which means that they cannot cover a topic in the necessary depth, but rather should be viewed as 'introductions'.

6. Would I hire somebody because he did courses at coursera?

It depends. If he is just posing with 'certificates', I would easily sort him out. If she has taken some really challenging courses there as a way of self-improvement, I would consider that a big advantage.

7. What are the best courses to take?

I guess most of the courses are excellent. I will review these in detail:
1. Foundations of Business Strategy (University of Virginia).
2. Financial Engineering and Risk Management (Columbia University).
3. Computational Finance and Financial Econometrics (Washington University)
4. Natural Language Processing (Columbia University)
5. Probabilistic Graphical Models (Stanford)
6. Start-up Engineering
7. Unpredictable? Randomness, Chance and Free Will (NU of Singapore).
... more to follow.
 

Thursday, May 2, 2013

Coursera review: Probabilistic Graphical Models

In this series of posts, I will share some experiences from the on-line learning platform coursera.org

Today: Probabilistic Graphical Models by DaphneKoller from Stanford University.


1. What is the course about?

Probabilistic Graphical Models (PGMs) is a framework of mathematical models. It spans methods such as Bayesian networks, Markov random fields, coding theory, and discrete data structures in used in modern computer science. The goal is to efficiently encode and manipulate high-dimensional probability distributions, often involving thousands of variables. These methods have been used in a wide range of applications, like: web search, medical and fault diagnosis, image understanding, reconstruction of biological networks, speech recognition, natural language processing, decoding of messages sent over a noisy communication channel, robot navigation, and many more.

The course includes video lecturing, discussion groups, test quizzes, final exam, and programming assignments. The programming assignments are needed only for students who wish to follow the 'advanced track'.

Level: graduate.

2. Who should use it?

People with very strong background in technical disciplines like mathematics, computer science or engineering. Students who are interested in understanding and building large-scale data analysis applications in complex, multidimensional environments.
 
3. Who should NOT use it?

Students with limited experience in algorithms and/or without strong technical background (including high-level math skills, not just programming).

4. What is the best thing about this course?

Excellent lecturer who explains clearly pretty complex notions through carefully selected examples. Presented are many links to real-world applications. Unlike other courses on coursera, this one is not 'watered down'.

5. What is the worst thing about this course?

The course claims that 'the average Stanford student needs between 15 and 20 hours of work per week to complete the course'. This might be true for some students. In my personal experience this statement was exaggerated, but nevertheless I needed about 7h/week (60% of which was spend on the programming assignments for the advanced track). For a working professional this might be too much.

Coursera review: Natural Language Processing

In this series of posts, I will share some experiences from the on-line learning platform coursera.org


Today: Natural Language Processing by Michael Collins from Columbia University.



1. What is the course about?

The course gives an introduction to the various types of modern statistical models used in natural language processing, including tagging, named entity recognition and machine translation. 

The course includes video lecturing, discussion groups, test quizzes and programming assignments.


Level: 4th year undergraduate or 1st year graduate.


2. Who should use it?

2.1. University students in technical disciplines like mathematics, computer science or engineering.

2.2. Working professionals with strong quantitative background as a form of relaxation.

3. Who should NOT use it?

3.1. Students with limited experience in programming, little interest in algorithms and/or without strong technical background.

4. What is the best thing about this course?

Interesting and challenging programming assignments, which help the student to understand clearly the theory. Excellent lecturer!

5. What is the worst thing about this course?

It might turn out to be time-consuming, in particular for people who have never been exposed to related topics like machine learning / AI.

Coursera review: computational finance and financial econometrics

In this series of posts, I will share some experiences from the on-line learning platform coursera.org


Today: Computational Finance and Financial Econometrics by Eric Zivot from University of Washington.



1. What is the course about?

The course gives an introduction to computational finance methods, mainly related to stock market investments.

The course includes video lecturing, discussion groups, test quizzes, programming assignments and a final exam (test quiz).


Level: 3rd year undergraduate.


2. Who should use it?

2.1. University students, in particular in technical disciplines like mathematics, computer science or engineering.

2.2. Working professionals in the finance sector or talented financial journalists as a form of relaxation.

3. Who should NOT use it?

3.1. Day-traders, chartists and any other dreamers thinking that the course would have any relevance to their investment strategies.

3.2. People who think that the course might have a significant impact on their career, both as a certification and as a know-how-set.

4. What is the best thing about this course?

You get exposed to R, the most powerful language for statistical data modelling.

5. What is the worst thing about this course?

The quality of video lectures is below the coursera standard. The first half of the course if absolutely boring for anybody with statistics knowledge. Overall, for industry professionals the course is easy to an extend that kills all pleasure of taking it. 

Coursera Review: Financial Engineering and Risk Management

In this series of posts, I will share some experiences from the on-line learning platform coursera.org


Today: Financial Engineering and Risk Management  by Martin Haugh, Garud Iyengar, Emanuel Derman from Columbia University.



1. What is the course about?

The course gives an introduction to financial engineering, mainly with respect to valuation of derivatives in a binomial model framework.

The course includes video lecturing, discussion groups and test quizzes.


Level: 4th year undergraduate or 1st year graduate.


2. Who should use it?

2.1. University students, in particular in technical disciplines like mathematics, computer science or engineering.

2.2. Working professionals in the finance sector or talented financial journalists as a form of relaxation.

3. Who should NOT use it?

3.1. Day-traders, chartists and any other dreamers thinking that the course would have any relevance to their investment strategies.

3.2. People who think that the course might have a significant impact on their career, both as a certification and as a know-how-set.

4. What is the best thing about this course?

During the quizzes you will develop a nice set of Excel sheets illustrating the main concepts of derivative valuation. Excellent lecturers with practical insight!

5. What is the worst thing about this course?

It is too lightweight for people with knowledge in risk management. At the same time it's too heavy-weight for people who see the topic for first time.

Coursera Review: Foundations of Business Strategy

In this series of posts, I will share some experiences from the on-line learning platform coursera.org


Today: Foundations of Business Strategy by prof. Michael J. Lenox from Virginia University / Darden School of Business.



1. What is the course about?

The course gives a basic tool-set for analysis of competitive positioning of a firm.

The course includes video lecturing, discussion groups, test quizzes, and a peer-reviewed final project (3-15 pages analysis of the competitive position of a firm, chosen, researched and written by each student).


Level: 2nd or 3rd year undergraduate.




2. Who should use it?

2.1. University students, in particular in disciplines like business administration or economics, but also as an enhancement to any other major.

2.2. Working professionals as a form of relaxation.

2.3. People being at or considering moving to any level management position, starting from product management up to the CxO level.

3. Who should NOT use it?

3.1. Entrepreneurs doing or planning to do a start-up might find it less useful for their purposes.

3.2. People who think that the course might have a significant impact on their career, both as a certification and as a know-how-set.

4. What is the best thing about this course?

It introduces a nice set of fairly visual tools which might be helpful for a presentation at management level. Experienced lecturer!

5. What is the worst thing about this course?


It touches very briefly upon interesting topics from macro-economics (e.g. Ricardian economics in the context of competitiveness) but unfortunately does not go into the necessary detail. As a mathematician I was so sorry to see a meaningful plot demonstrating some macroeconomic equilibrium shown for a few minutes and disappearing without being properly defined and explained.