Allow me to write this after receiving enough reference requests.
Overall, I like to help you as much as possible.
If you are applying for a MS/course program, the reference letters probably are not very important, as the top US schools (to my knowledge) mainly look at your statistics, like GPA, ranking, GRE, etc.
To put it more bluntly, MS program is a source of revenue for them.
For this, all you need is to have obtained top grades in the courses I have taught.
However, I can only comment your specific course performance but not extrapolate, e.g. from basic programming (a class you took with me) to machine learning (a class you did not take with me).
If you are applying for a PhD/research program, you need to have at least some good publications. Any decent professors/researchers know that good grades do not imply good research potential. (I am not aware of any rigorous study, but I think the two are weakly positively correlated at best). Thus, I will write letters only for those who have published top research papers or built good industry products with me, as otherwise the recommendation is likely moot.
Comments Off on References for MS/PhD applications
When they just arrived they thought it tough to publish at least one first-authored SIGGRAPH paper before graduation.
Now they are hitting the job market and found out that some topic research lab (not to be named but this is no secret) requires at least 5 first-authored SIGGRAPH papers.
Birds of a feather flock together.
Your opinions of others often reflect more of who you are than who they are.
I would like to thank those who (unintentionally) help me filtering away unqualified candidates; you might be even more effective than what the good folks could do.
Comments Off on Qualification filter
When I was a PhD student, having 1 SIGGRAPH paper meant graduation, and 2+ for a top research job.
Now, having 1 SIGGRAPH paper meant admission into a top PhD program, 2+ for graduation, and 3+ for a top research job.
(20+ for a tenured professor or partner researcher, but few of you need to worry about this yet.)
Anyone who (still) thinks my standard is too high: feel free talk to Jun Xing, my first HKU advisee, about his current experiences in internship and job hunting.
Comments Off on Performance inflation
Hi Jun, Qi, and mamba:
You know what kinds of students I am looking for.
Some of the applicants might contact you about my advising style.
Please help me screen away those who are not suitable (e.g. sharing your painful SIGGRAPH experiences should be a good start).
They trust your words more than mine.
It is to your advantage to be surrounded by top players.
And the fewer students I have, the more time I can spend on each of you.
Comments Off on Asking for a favor
Nowadays a quick way to filter a job/school application is to see whether and how it says the candidate wants to do machine learning.
(Some neural network probably already existed precisely for this.)
Machine learning by itself is not the problem (quite on the contrary).
The problem is whether you can even form your own independent opinions.
When something (investment, technology, or research field) becomes hot it is already too late to bandwagon.
Those pioneers you see today started (and stuck to) their stuff when it is not yet hot.
Stick with your passion, belief, and opinion might not lead to success, but at least you can have fun, face less competition, and success/fail in your own style.
And if you are smart and creative enough you can have the cake and eat it.
Say your expertise and/or interests are about user interface design. But you also want to do some machine learning like everyone else.
You can switch field, and compete with a lot of smart people who have more passion and knowledge.
Or you can stick with user interfaces, and use machine learning to make them better. You can pick up something new without ditching what you already have.
Comments Off on Hypes
One common advice on research is to have a coherent theme among our papers. I heard this from a bigwig around 2003 after getting my PhD.
This is one of these advices that I agree in principle but have violated in practice.
Yes, coherence can help recognition from the community, especially when one enters a new field.
However, I am not sure if this should be intentionally aimed for. Unless you are extremely smart and versatile, you are likely end up doing related stuff without even trying.
There is this implicit force that drags us towards similar, and thus incremental, ideas. We should fight against this force, not follow it.
So, just do whatever you like. You will have more fun and more likely to produce novel stuff which, even if lacks coherence, beats being incremental.
Comments Off on Research coherence
Using slides is a popular way to give presentations. I am not sure if it is the best way, but things can go very badly if done in the wrong way.
Take a look at Jim Blinn’s post about giving presentations.
Below are some quick high level suggestions. (I plan to refine this post later.)
Aim for simplicity and minimalism.
The slides are for conveying information to your audience, not serving as memo for the speaker.
Use intuitive pictures, illustrations, and animations, instead of texts and (worse) equations.
If you find yourself worrying about typography, it is a sign of too much texts.
No sentence should run over one line.
Rid of visual clutters like bullet points.
Gratuitous colors and unnecessary font variations tend to confuse people.
Comments Off on How to design talk slides
Talks are usually easier to understand than the corresponding papers. To get accepted, papers need to be written in a way that look formal and rigorous, but not necessarily easy to understand. However, when authors present their (accepted) papers, they tend to cut the chase and talk straight.
In the past, people have to attend conferences for the talks.
Nowadays, everyone can easily share their talk slides online.
Better yet, record and share your presentations as videos (e.g. via PowerPoint). This can be done during practices or official presentation.
You do practice your talks, right? So why not record during your rehearsals, so that you can review now and share later.
Recording in official presentation might be trickier, e.g. the conference may prefer presenters using a shared machine and the recording might disrupt your presentation, but can be worth a try.
I have not done this for my own talks, but realized it can be a good idea after watching a few recorded talks online. I really appreciate the efforts from the authors, and plan to do so for my future talks.
Comments Off on Recording and sharing presentation
A collaborator asked me about time management, in particular how to toggle between multiple projects.
I am not sure if I am any good at time management or more general if I am a normal person, but here is what I do.
Rule 1: do not block your collaborators.
Rule 2: do whatever you like otherwise.
During my NVIDIA days I belonged to a tight-knit group of around 40 to 50 GPU architects. GPU design was (and likely still is) tremendously complex, so anyone could potentially block everyone. When that happened, ATI got a chance to kill NVIDIA, and my colleagues got a chance to kill me. That is how and where I learned rule 1.
In general, prioritize your collaborators over yourself. It may sound counter-intuitive, but making others happier and more productive will make the entire team, including yourself, happier and more productive as well.
Rule 2 reflects my personal style: after making my collaborators happy, I make myself happy. I cannot be productive otherwise, and life is too short anyway. I believe most people over-plan, especially for creative and non-deterministic tasks like research. This is one main reason why I never schedule regular meetings or daily routines. If you have a hard time deciding what to do first, they are probably equally important, so just pick one and do it. Trust your instincts, and you will gradually learn how best to schedule your time (this could help, which I also learned from NVIDIA).
I am not familiar with other time management techniques like Pomodoro. I just took a quick look at that and found myself violating it badly. For example, if a distraction pops into my head, I follow it, to allow serendipitous encounters (unless I am right in front of a major deadline). My distraction stack is probably around 4 layers deep, i.e. I could recurse around 4 layers of inception without getting lost.
Comments Off on About time management
It is usually not too hard for smart kids to perform well in schools; just excel in what you are told to do, such as taking courses.
This is a deterministic process with well-defined goals and tasks that reward smartness and hard working.
However, real world is chaotic and ambiguous. You have to figure out what to do, with shifting targets and ever-changing environments.
This is why school performance does not directly translate to real-life performance: the required mentality and skills are not the same.
This is also why being a good student might not be a good thing for you. You are so used to this deterministic input-output process that you might be very frustrated by the non-deterministic nature of the real world, when starting your first job or research project.
In contrast, not-so good students might adapt better to the real world, because they already have enough failure experiences and are not yet cast into conformity.
I was lucky to be a student who was considered good in performance and bad in behavior.