Confessions of a researchaholic

2013-02-14

Open mentoring

Filed under: Real — liyiwei @ 12:23 pm
Tags: , ,

If you want to strengthen your credentials for school/job applications while having fun learning about computer science research without any additional costs or hassles, I can help you.

See below (and my advising style) for details, and let me know if you have questions or comments.
Please also help spread the news; we are making this world a better place.

Why

It is becoming more and more competitive to get to a good job or school. You need a glistering resume. And passive measures like grades, ranks, and standard test scores are no longer sufficient because they do not reflect the active skills which are crucial for today’s highly dynamic and creative job functions.

School application example: 20 years ago if you are a top 5 percent student from a number 1 department/school in your specific country/region with good GRE/TOEFL scores (all passive measures), you can probably get into MIT/Berkeley/Stanford PhD program without much problem.
Today, you will likely not make the cut without impressive active measures such as publications and recommendations from top people in your field.

Job application example: 20 years ago a Stanford PhD without any publication can probably beat a Tsinghua PhD with strong publications in getting a high tech job in the US. Today will be the exact opposite.

In a nutshell, active measures are gaining importance over passive measures. But they are also harder to come by entirely on your own.
This is why people are doing all these internships and school projects. The question is: how good are your internship/project mentors, and how much credibility their recommendations carry? If you plan to spend time on these, better pick a good mentor.

What

I have been mentoring and collaborating with many students and junior researchers for a while. I have this unique asynchronous style that is not only very effective (judging by the publication records and responses from my collaborators) but also very scalable (absolutely no resource constraint except our passion and commitment).

If you do well, you will get strong publications and strong letters of recommendation from me. I am well connected to top schools, companies, and recruiters. You will also have a lot of fun with your projects.
If you do not do well, you have nothing to lose, as long as you do not list me as a reference.

How

I work with you asynchronously through svn paper drafts and Google sites. It is up to you to decide when and where you work. The amount of time I spend on you is entirely proportional to your productivity. I seem to have this uncanny ability to remotely read human minds (and sense human emotions) more effectively than ordinary people can face-to-face (the origin of my “Jedi” nickname).

You start by telling me what interests you, and I brainstorm with you to find a good project direction.

I then pick a warm-up project, usually reproducing a known piece of work in your field of interest, such as writing a ray tracer (my favorite pet project for rendering) or implementing your favorite conference/journal paper.

Not everyone will survive this warm-up stage. But if you do, we move on to a creative project, aimed for a real conference or journal publication. You will have a chance to learn everything that ever needs to be learned about doing research and publishing a paper. I can claim this because I have several single-authored papers to prove the completeness of my skill-set. I also have papers with different numbers of authors to prove my collaboration and advising skills.

Who

I am most experienced with senior-undergrad + graduate students as well as junior lab researchers, but I do not have any hard constraints. It really depends on how we feel about each other.

I particularly like to help those who lack proper guidance; there appears to be many irresponsible/incompetent mentors out there, so I consider myself doing a useful social service.

If you want to become my internal student, you will need to go through this process anyway.

When

As indicated above, we can start anytime you like.

However, you might want to time your school/job applications. For example, it is not sensible to start working with me a month before your school/job application deadline, because I will have little to say about you. A good rule of thumb is to contact me at least a year ahead.

Where

As indicated above, you can work anywhere you like.
I have a shared online work-space for you to meet others, as well as channels for individual discussions.

I might be able to arrange a few internship quotas, but those are already very competitive, and frankly they make little (if any) difference for my mentoring style.

2013-01-25

Teaching

Filed under: Real — liyiwei @ 7:23 pm
Tags: , ,

My attitude towards teaching could be reflected by how I would really like to write my teaching statement (for faculty job applications).
There will be only one sentence: “smart enough students can pick up everything themselves”.
(Note1: I know this is entirely doable because I pulled this off since I was 10 years old, and I did not consider myself to be very smart.)
(Note2: Of course this is not how I really wrote my teaching statement.)

With this backdrop, it should not be a surprise that I have never blogged a single entry about teaching, at least not mine. I do not even have teaching as a tag word.

I have considered teaching as a chore rather than enjoyment (unlike research), and my basic stance has not changed too much. (One main reason why I think it is better to start a research career as an industry lab scientist rather than a university professor.) But not until I really taught full-semester classes, especially large ones, did I start to appreciate teaching can be a fun thing to do, for two main reasons.

First, it can actually inspire my research ideas.

Second, and probably more important, teaching provides a great chance for massive mind reading and human studying, with moral justification for effective teaching. It is even more fun and challenging than reading individuals, one of my most favorite pastimes.

More posts to follow.

2013-01-07

Sharing code and data

Filed under: Real — liyiwei @ 1:08 am
Tags: ,

It is usually a very good idea to share the code and data along with our published papers. This will make it easier for others to test, understand, reproduce, and compare against our methods, which in turn can make us more popular, our papers more widely cited, and our technology more likely to be adopted by the industry and turned into real products.
Code and data repositories are also an important part of evaluating job applications.

(I have open sourced most of my first/single-authored projects, except my first 2 SIGGRAPH papers for which I could no longer find the code in the school server, which I greatly regret.)

Ideally the code should be in high quality, but even if not, sharing it can let others have a chance to improve the code (and motivate us to write good code).
It is fair to say that the code and data are no less important than the paper.

Some things to watch out for include institutional and legal constraints, such as trade secrets, copyrights, and patents, and not yet published future ideas.
These can be planned and managed via different git branches (e.g., a public branch that is gradually merged from a private branch).

2012-12-31

How to design demos

Filed under: Real — liyiwei @ 9:51 pm
Tags: ,

In a nutshell, a demo should properly demonstrate technical aspects with sufficient artistic appeals.

The technical part is usually more important, and can suffice alone for many science and engineering disciplines. However, the artistic part is also very important for graphics and HCI, or any fields which involve direct human perception and consumption.

Demos usually take a lot of time and efforts, on top of the usual workload in ideation, writing, algorithm, implementation, and experimentation. And whether you like it or not, a solid and novel algorithm cannot be adequately assessed or appreciated by the readers if it is not demonstrated through proper demos.

Thus, designing demos is kind of an art. Below are recent suggestions from Sylvain Lefebvre which I have found to be excellent.


A guideline that worked fine for me is to consider whether 1) the result demonstrates the technique properly and 2) the result looks just good enough that it appears useful; in particular we want to avoid people think that the example is contrived to only show the advantage of our approach.

The problem is that 1) and 2) sometimes compete with each other (e.g. a fantastic rendering possibly making it hard to properly see the motion, etc.). Also we do not want to spend too much time on 2), only enough that people will think that it is convincing.

2012-12-27

Massive open online courses

Filed under: Real — liyiwei @ 1:40 am
Tags: , ,

Finally, technology is turning its head towards one of the most important and yet archaic aspects of our civilization: education. I have been trying out these MOOC (e.g. Udacity and Coursera), and found them fascinating. I can understand why Sebastian Thrun decided to quit his tenured Stanford faculty job to startup Udacity.

This is going to totally shake the entire education sector. Schools not on the efficient frontier, especially those primarily focusing on teaching rather than research, are in the real danger of extinction.
Pretty soon, students will start to ponder between getting course certificates from Stanford versus getting real degrees from lesser schools.

The technology is no longer the issue. Stanford, if it wants, can already dish out an infinite number CS master degrees annually. The real question is prestige; Stanford is valued precisely because it is a scarce resource. How MOOC and prestige will play out remains an interesting event to watch for.



2012-12-11

A gastronomic study of PhD defense

Filed under: Real — liyiwei @ 6:29 am
Tags: ,

🙂

HKU (Hong Kong)

No food or drink served. The candidate gives the talk, the audience asks questions, and the jury members discuss the final fate behind the door. The entire thing is over in about an hour.

Stanford (USA)

The reception food quality determines whether you pass, but do not serve coffee, as it can make jury members unnecessarily aggressive. The rest is similar to above.

INRIA (France)

The jury members meet for lunch in a nice restaurant (French, of course). The talk, QA, and jury discussion parts are similar to above. After that, a fancy reception is served (Tunisian pastry in my case). And if you, as a jury member, do not ask for trouble during the defense, you get invited to dinner in an even nicer restaurant (French, course). The entire thing takes about 11 hours.

2012-12-02

Novelty

Filed under: Imaginary,Real — liyiwei @ 5:55 pm
Tags: ,

We like the novelty of giving up what we know, and we like the novelty of coming to know something we did not know. Otherwise, we would just hold on to what we have, and that's not very interesting. - Jasper Johns

2012-11-21

Flesh UI

Filed under: Real — liyiwei @ 2:52 am
Tags: ,

I have a primary bank in Asia and a primary bank in the US.

The online banking of that US bank is very easy to use; the UI design is likely good enough for a textbook example.

In contrast, online banking for that Asia bank is extremely difficult to use (even for a computer scientist); the UI design is so bad that it can also serve as a textbook example.

I have been wondering how that Asia bank could have survived its customer complains about UI. Today, I found out the answer.

I needed to conduct a transaction that (to the extent I can figure out through that UI jungle) cannot be done online. So I went to a local branch. The (human) bank clerks were so efficient that the entire process probably took less time than doing it online by myself.

So I guess most people in Asia simply do banking with real humans. It is not only because human labors are more abundant in Asia, but also that Asians have more inherent distrusts for online banking.

Lesson learned: a bad UI needs fix only if used.

2012-11-10

How I work with each student

Filed under: Real — liyiwei @ 11:34 am
Tags: ,

This figure illustrates how I collaborate with each student.

There are several main layers of a research project: idea creation, algorithm design, system implementation, and experimentation + results production. They are all wrapped up and glued together by presentation, which includes paper writing, video production, and (conference) talk.

For a beginner student, I would expect you to completely take charge of the implementation and production (including video), given the main ideas and algorithms as described in a paper draft (as part of the presentation). We could discuss anything you like, but you should be able to handle all the implementation and production details (e.g. I am not supposed to look into your code). Otherwise, you are not ready for research.

As you progress, I would expect you to be able to take charge more tasks at the higher layers. Idea creation is the most difficult and only a few can pull that off, but my experience is that a decent senior student can at least take part in algorithm design and paper writing.

« Previous Page — Next Page »

Theme: Rubric. Get a free blog at WordPress.com