a thank you note

Alhamdulellah, I was recently awarded the Google PhD fellowship award. I'd like to take this opportunity to express my gratitude to people who invested a significant time and effort in my career. Ordered by the date we met, and mentioning just one of the many things they did for me, I dedicate this award to:

• Mama, Mabroka Ammar, who taught me empathy.
• Baba, AbdelMottaleb Ammar, who taught me persistence.
• Khaled Ammar, my brother, who made me a computer engineer.
• Amany Shehata, my wife, who sacrificed her best interests so that I can pursue my passion.
• Kareem Darwish, who transformed me into a curious researcher.
• Noah A. Smith, who taught me writing and critical thinking.
• Chris Dyer, who gave me confidence and great ideas.

While we are at it, and since this is likely to be the last award I get (I don't know how I got it in the first place :D), I'd like to also say thank you to each of: my daughter Salma Ammar who makes me smile; my role model ElDesouky Ammar; my loving siblings and relatives Khaled Ammar, Samia Ammar, Walaa Ammar, Mohamed Amir, Mohamed Ghazy, Hesham Ghazy, Mohamed Ammar, Ahmed Ammar, Ibrahim Ammar, Abdelrahman Ammar, Hassan Ammar, Ali ElTamalawi; my high school math teacher Mostafa ElShantoury; my professors Larry Wasserman, Ryan Tibshirani, Soheir Bassiouny, Lori Levin, Tom Mitchell, Mostafa Youssif, Salah Selim, Mohamed Nazieh, Mohamed Ismail; people who wrote me recommended letters D. Sculley, David McClosky, Kuzman Ganchev, Motaz ElSabban, Nayer Wanas, Tarek ElAbbady, Soheir Bassiouny, Kareem Darwish, Ayman Kaheel; my helpful (ex-)colleagues and ex-managers Ahmed Ramadan, Ahmed AbdulHamid, Mahmoud Refaat, Ahmed Morsy, Ahmed Sabry, Hani Salem, Kareem Darwish, Hamdy Khalil, Mohamed Ali, Nayer Wanas, Ayman Kaheel, Tarek ElAbbady, Ahmed ElShimi, Federico Garcea, Takako Aikawa, Arul Menezes, Kristina Toutanova, Hoifung Poon, Chris Quirk, Vikram Dendi, D. Sculley, Brendan O'Connor, Ali ElKahki, Walid Magdy, Swabha Swayamdipta, Lingpeng Kong, Chu-Cheng Lin, Dani Yogatama, Ankur Parikh, Daniel Golovin, Julian Grady, Lirida, Ahmed Hefny, Manaal Faruqui, Shay Cohen, Kevin Gimpel, Andre Martins, Avneesh Saluja, Hany Hassan, David Bamman, Sam Thomson, Tarek Hefny, Austin Matthews, Dipanjan Das, Jesse Dodge, Matt Gardner, Mohammad Gowayyed, David Pellow, Avi Dubey, Yan Chuan Sim, Eva Schlinger, Victor Chahuneau; my friends who made life more tolerable Ashraf Gamal, Mohamed Darwish, Ahmed Hefny, Ahmed AbouTaleb, Mohamed ElShatnawy, Taher Gomaa, Ahmed Mounir, Islam Gohar, Ada Iosif, Ahmed Abdelgawad, Hatem Kassab, Ezzeldeen Qasim, Osama Sharkas, Abdalla Musmar, Hany Hassan, Vikram Dendi, Ahmed ElBeshir, Justin Chiu, Yubin Kim, Prasanna Kumar, Meghana Kshirsagar, Wang Ling, Subhodeep Moitra, Leah Nicolich-Henkin, Kenneth Huang, Reyyan Yeniterzi, Shoou-I Yu, Amr ElAlfy, Mohamed Diab; and, of course, my lifelong friends Hesham Zakareyya, Mahmoud Refaat, Mohamed Saleh, Haitham Gabr, Amr Ebaid, Hani Desouky Ammar, Hany Abdelrahman Ammar, Mohamed Ghareeb, Mohamed Ibrahim, Mostafa ElMarasi, Mohamed Mokhtar.

P.S. You may find this hard to believe, but I actually am indebted to each and every one in this long list. I'm just worried I may have forgotten someone.

The pomodoro technique

If you are the kind of person who easily gets distracted with emails, phone calls and minor tasks, you may find the Pomodoro Technique useful.

technique: 

1. get a count-down timer
2. pick a task
3. set the timer to 25 minutes (aka. a pomodoro)
4. work on the task, with no interruptions until the pomodoro ends.
5. add an(other) 'X' mark next to the task (even if it's not completed).
6. take a short break (3-5 minutes). after four pomodoros, take a longer break.
7. while task not complete, go to step 3.
8. after a task is completed, count the number of X's to get a sense of the time spent on it.

justification:

• time management should be simple
• frequent breaks improve mental agility [ref which I don't really understand :-)]
• increased awareness of the time spent
• reduce the time spent on distractions

caveats:

I've been following the pomodoro technique for a little over a month now, and I think it did help me spend the time more efficiently (at least when I'm serious about it). I'd like to highlight a few problems and how I attempted to fix them:

problem: just because you shouldn't be distracted in the middle of a pomodoro doesn't mean you won't. 

solution: every time I get unnecessarily distracted, I punish myself by resetting the pomodoro.

problem: the short break is too short.

solution: make it a little longer. I found 10 minutes to be a reasonable time.

problem: when tasks take too many pomodoros, I lose any sense of achievement.

solution: spend the first couple minutes of a pomodoro thinking about the focus in this pomodoros. then, spend the last couple minutes of a pomodoro documenting what you have accomplished (regardless of how little/insignificant it was).

complementary tools:

While I found this technique to be quite effective, it doesn't address all time management problems. I use rememberthemilk and github issues to track my tasks and prioritize. I keep all my time-constrained commitments on the calendar to make sure I'm not double booked. ...etc

Choosing an advisor

Recently, I've been in touch with many students in the process of finding a PhD advisor. While it's not impossible to change advisors, it usually comes at the cost of more time spent in the PhD program. Here's a list of things you can do to improve your chances of finding a good match:

For each potential advisor X, do the following:

1. Make sure you don't have a problem with X's research interests.
2. Make sure you can get along with X.
3. Talk to X's current students about X's pros and cons. Ask clever questions since many students refrain from discussing the downside of their advisors with a stranger. Don't forget to ask if X is too (dis)engaged with his students. This step also helps you find out more about your potential academic siblings.
4. Find out what project would fund you. Is it a long-term or a short-term project. If it's a long-term project, make sure you like it.
5. Find out if X has a tenure, in tenure track, or neither. Non-tenure (track) faculty are more inclined to leave their academic institution and find a better position somewhere else, which is -usually- a bad thing for their students. Tenure-track faculty are more inclined to work harder, and tend to be more demanding (which can be good or bad, depending on who you are). 
6. Collect some statistics about X's research group. How many students does X advise? A very large number (e.g. fifteen) usually means X won't have time to interact with you, and a very small number (e.g. zero/one) usually means X has problem with funding (or a new hire). Find out what's the average lifetime of X's students in the PhD program; the variance is usually small. 
7. Find out if X collaborates with other academics (inside or outside the department/school). More collaborators usually mean a well-connected advisor, which can be very useful for your marketability.
8. Find out where does X's graduates go.
9. Meet with X. Ask smart questions. Be ready to talk about your previous work.

If feasible, do a research project (e.g. in a lab or a regular course) with your preferred advisor before committing to be his student. You might change your opinion after working with him/her for a while.

The evolution of intellectual freedom

I thought I would joke about this before it becomes a bitter reality!

ref: PhD Comics

The seven secrets of highly successful PhD students

Vitae is an organisation championing the personal, professional and career development of doctoral researchers and research staff in higher education institutions and research institutes.

Their website has a number of useful resources for graduate students. One of them is a document titled "The seven secrets of highly successful PhD students". Here's my summary:

1. Keep in touch with your supervisors.
2. Write often.
3. Be realistic about your contribution and yourself (being different than your PhD).
4. Avoid distractions.
5. Treat PhD as a job.
6. Ask for help.
7. Understand it's all about persistence.

Many thanks to Dina Said who let me know about this resource.

[metapost] Apology

I apologize for being away for a while. A number of significant events in my personal and professional life kept me busy for the past few months.

First and foremost, as of today, I'm the proud father of a sweet baby girl. She got beautiful eyes, minute fingers, loud cries, lovely smile and she has got a name: Salma. I'm planning to make friends with her, but she doesn't seem to be an easy girl. She seems to know her mother very well, though.

Another significant event that occurred to me of late is the Egyptian revolution. I didn't feel so much pride being Egyptian until that day! Although I was thousands of miles away from the Tahrir Square (center of the Egyptian revolution), I was honored to participate with my voice to form the international pressure on Mubarak's regime.

On the professional level, I have recently joined the Machine Translation team in Microsoft Research, Redmond. This team is a blend of talented researchers and engineers working together to push the boundaries in this field AND shipping that to create a competitive solution. I'm definitely proud of belonging to this team.

The last piece of news I'd like to share with you today is that I got offers from CMU, UMD, UBC and Purdue graduate schools. I'm planning to visit a couple of them before deciding where I'm gonna spend the next 5 years of my life.

I'm back!

How to be on top of your field

Most computer science fields are fast moving. There’s a lot going on all the time at universities and research labs. Hundreds of publications appear every year through the few major conferences and journals of a field, not to mention venues with lower ranking. It is particularly important to keep a close eye on the advances going on in your field because research is –usually- incremental. You start where others left off. If you don’t know the recent findings of other scientists in the field, you will have hard time figuring out what topics are important, what approaches are popular, what the state of the art solving a particular problem is, etc.


Problem is: we spend a lot of time working on a particular project, taking care of our day-to-day job responsibilities, and trying to meet deadlines. It’s not uncommon for such activities to take up all the time we have, leaving no time for expanding our knowledge learning about recent findings in the field.

One way to solve this problem is to attend related major conferences. In addition to learning about new advances presented in the conference, you will have the opportunity to mingle with other researchers working on similar/related problems which may lead to collaboration projects. Unfortunately, this is not always a viable solution. Who got the bandwidth and resources to attend all related conferences!

Another solution that’s more affordable for research groups in universities as well as research labs was described by Prof. Azer Bestavros as follows:

• Each member in the group maintains a list of interesting papers s/he would like to read, covering last year’s proceedings of the major conferences and journals in the field.
• Schedule a recurrent 15-minutes meeting (frequency depends on the group size) in which one of the group members gives an overview on a paper in his list.

For example, a group of size 10 may choose to hold this meeting every other day. This way, every member gives 0.5 talk per week and listens to 2.5 talks per week. Good deal, huh?

Personal statement

This post is the fifth featuring Prof. Mor Harchol-Balter's talk advising people applying to PhD programs in computer science or related areas.
[1st episode - 2nd episode - 3rd episode - 4th episode - full article]

It’s misleading that the personal statement is called a “personal” statement, since what admission committees are really looking for is a research statement. What admission committees want is a statement about what research you have done, what research you hope to do, and why you like research.

Here’s a template if you need one:

i. First paragraph – Describe the general areas of research that interest you and why. (This is helpful for a committee to determine which professors should read your application.)
ii. Second paragraph and Third paragraph – Descibe some research projects that you worked on. Tell us what you found, what you learned, what approaches you tried. It’s fine to say that you were unable to prove what you wanted or to solve your problem.
iii. Fourth paragraph – Tell us why you feel you need a Ph.D.. Look back to section 2 and explain what in there appealed to you.
iv. Fifth paragraph – Tell us why you want to come to CMU.Whom might you like to work with? What papers have you looked at from CMU that you enjoyed reading? What will CMU teach you?

It’s important to realize that the research statement is not a commitment to do research in that area. A third of all applicants end up working in an area different from that which they described on their research statement.

Here are the common mistakes that half of our applicants make:

• The grade regurgitator – “In my high school, I was ranked Number 1. Then I got a perfect score on my college entrance exams. Then I competed in a statewide math competition and I was the best. Then I competed in a national programming competition and I was 5th. In college, my GPA was 3.95 out of 4.0. For these reasons, I believe I will do well in your graduate department.”
What’s wrong with this? This portion of the essay is a waste of space. Awards are certainly relevant, however any award you won should be listed on a separate piece of paper which is titled “Awards and Honors” and which you can include with your application. There is no reason to tell us all this in your essay. It will only piss-off the people reviewing your application because they already read all this information earlier in your application and they now want to hear about research only.

• The boy genius – “When I was born, my mother gave me a glass ball to play with. I would lay and look at the prisms of light shining through my ball. At age 3, my father brought home our first computer and I disassembled it and then put it back together. It was then that I knew I wanted to become a computer scientist. By age 5, I had taken apart every appliance in our house. At age 6, I became a chess whiz ....”
What’s wrong with this? We simply don’t care what you did as a child, and we don’t believe you either. You’d be surprised how many applications from Einstein-wanna-be’s we get. If you really think this is relevant, put the important facts on a separate sheet of paper, and include it in your application. It’s best if your essay can stick with stuff you did in college and later.

Failure is an orphan

An eye opener! That's the least to say about this inspiring article for Prof. Stan Szpakowicz in Computational Linguistics journal.

According to the article, one project out of a hundred produces results that justify the investment. However, we tend to count ourselves among the 1%. That's because we need to show we outperformed others in order to publish more in order to have a thriving career in research. That's a compelling reason for us NOT to invest more time on something that produced negative results.

But WAIT. "Suppose you have set up an experiment carefully and in good faith, but still it comes up short. That’s not a positive outcome. Maybe your intuition has let you down. Maybe this cannot work. Wait, maybe you can prove that it cannot work?" THAT would be a useful outcome. But you need to get it published.

The problem boils down to peer reviews which aggressively reject failures. "A forthright admission of the inferiority of one’s results—despite the integrity or novelty
of the work—is a kiss of death: no publication. There must be improvement ... conformance to reviewers’ expectations is an asset. Indeed, we write so they are likely to accept". But we're talking about ourselves. WE are the reviewers. If we -the authors- insist on striving for success and run away from every negative result, then we -the reviewers- will check papers for signs of success.

An experiment carefully thought out, a systematic procedure, an honest evaluation—these are the ingredients of good science. It is not mandatory for the results to be positive, though it certainly lifts one up if they are. In areas where empirical methods dominate (e.g. computational linguistics) people try things which fail at the experimental stage. This may be due to lack of rigor, but often there are deeper, unexpected, and intriguing reasons. We can learn a lot if we analyze scientifically why an intuitive and plausible experiment did not work. Then again, to know what leads to dead ends in research surely can warn others off paths which take us nowhere. Simply put, a negative result can be a useful lesson.

Q: Philosophy aside, I want to publish a serious, worthwhile negative result I've obtained. Where to go today?

A: Here's a non-comprehensive list in different disciplines

• Journal of Interesting Negative Results in NLP and ML (http://www.jinr.org/)
• Journal of Negative Results in BioMedicine (http://www.jnrbm.com/)
• Journal of Negative Results – Ecology and Evolutionary Biology (http://www.jnr-eeb.org/)
• Journal of Articles in Support of the Null Hypothesis (http://www.jasnh.com/)

The illustrated guide to a PhD

Here's how Prof. Matt Might (University of Utah) explains to fresh PhD students what a PhD is:

Imagine a circle that contains all of human knowledge:
















By the time you finish elementary school, you know a little:
















By the time you finish high school, you know a bit more:
















With a bachelor's degree, you gain a specialty:
















A master's degree deepens that specialty:
















Reading research papers takes you to the edge of human knowledge:
















Once you're at the boundary, you focus:
















You push at the boundary for a few years:
















Until one day, the boundary gives way:
















And, that dent you've made is called a Ph.D.:
















Of course, the world looks different to you now:
















So, don't forget the bigger picture:
















Keep pushing.

Taken from http://matt.might.net/articles/phd-school-in-pictures/ - licensed under the Creative Commons Attribution-NonCommercial 2.5 License.

Research competitions

Research competitions are designed to accelerate research on a particular topic. The entity which organizes a competition usually have a problem and wants to encourage researchers to find solutions of its problem.

Examples:

• For example, Netflix, a popular US company which provides flat-rate DVD rentals and video streaming services, put a $1,000,000 prize for those who come up with the best collaborative filtering algorithm to predict user ratings for films based on previous ratings.
• Text Analysis Conference (TAC) is a series of evaluation workshops organized by NIST to encourage research in Natural Language Processing.
• Text REtrieval Conference (TREC) is a series of evaluation workshops organized by NIST to encourage research in Information Retrieval.
• OpenMT is yet another evaluation series organized by NIST to encourage research in machine translation technologies.
• Speaker Recognition Evaluation (SRE) is NIST's workshop to encourage research in speaker recognition.

Why should you participate?

• Data: Organizers of a research competition provide participants with scarce data resources for free so that they can compete. It is very expensive to collect the data yourself. Sometimes, you can subscribe to get the data for (huge) fees, but even then, data catalogs are not made available until many years after the competition was held.
• Evaluation: Normally, you need to prove your novel technique performs better than state-of-the-art techniques that handle the same problem as yours. First, you need to decide which other techniques you should compare to, which is not always an easy task. Then, you try to obtain the same data set used in their publication so that your results are comparable. Soon you find out they were not using standard data for training or testing. So, you decide to run the other technique on your data, but you can't find a readily available implementation of it. So, you have to implement it yourself. After all, the comparison may not be accurate because there are usually tons of details not mentioned in publications which make a big difference in results. When you participate in a research competition, you don't have to worry about all this painful overhead.
• Exposure: Normally, when you do something great, no guarantee people will listen to you. Most prestigious conferences, for example, reject high quality papers because they have limitations on the number of papers they may accept. When you participate in such competitions and produce great results compared to other participants, they will listen and learn from what you did.
• Publications: This is related to the previous point. Most competitions provide a good publication venue for participants to explain their systems and results.

You and your research

This is the title of a talk given by Richard Hamming in 1986. In this talk, Hamming was trying to address the question "Why do so few scientists make significant contributions and so many are forgotten in the long run?".

In the latest issue (Summer 2010) of ACM's XRDS magazine, Daniel Lemire reflected on that talk in an article titled "Marketing Your Ideas". If you don't have time to read the full transcript of Hamming's talk, you may want to have a look at Daniel's article. Hmm.. If you don't have time to read the transcript, you probably don't have time to read Daniel's article. Let me give an extractive summary of the latter:

1. Take your time

"Young scientists tend to rush their presentations. They work four months to a year on a project, yet they wait until the last minute before writing their paper and rehearsing their presentation—when they rehearse it at all."

"What about reports and research papers? Rushing their publication is trading quality for quantity. It is an unfortunate trade, as there is a glut of poor research papers, and too few high quality ones. Continuous writing, editing, and rehearsal should be an integral part of your activities."


2. Reach out to your audience

"Scientists and engineers are most successful when their work is most available. "

"But posting your content and giving talks is hardly enough."

"If you want people to attend your talks, make sure your title tells them why they should attend. Think about your audience. They want to know whether they should continue reading your paper or come to your talk. Convince them that you have something remarkable to tell them. Avoid jargon, acronyms, and long sentences."

"Do not underestimate email. It is the most powerful medium at your disposal. Yet, you have to use it wisely. To get famous people to read your emails, study their work. Show appreciation for their results. Think of reasons why they might find your question or proposal interesting."

Research interests change

One of the common mistakes scholars make is to assume a researcher/professor is still interested in a topic he/she has been working on 5 years ago! This usually happens when a student is looking for a supervisor/collaborator with a specific interest. The student finds a good paper on that topic and sends to the author.

As a matter of fact, researchers change their scientific interests over time. This might happen for several reasons. Government and/or industries fund research on a topic when a need arises, then turn down the fund when the need decays. Also, believe it or not, some research problems are eventually solved! Sometimes, researchers also change their focus moving to a new position to align with the research direction of the employer (be it a university or a research lab).

Should I get a Ph.D.?

This post is the fourth episode featuring Prof. Mor Harchol-Balter's talk advising people applying to PhD programs in computer science or related areas.
[first episode - second episode - third episode - full article]


Here are some things to keep in mind when making this decision:

1. A Ph.D. is not for everyone!
2. A Ph.D requires 6 years on average. The opportunity cost is high.
3. Do not even think of applying for a Ph.D. if you have not tried research and/or teaching and found that you like at least one of those. (Note: the Ph.D. program will require mostly research, not teaching, but a love of teaching may help motivate you to get through, so that you can go on to be a teacher. I have seen many examples of this.)
4. A Ph.D. requires a particular type of personality. You need to be someone who is obsessed with figuring out a problem. You need to have tremendous perseverence and be capable of hard work. You need to be willing to do whatever it takes to solve your problem (e.g., take 5 math classes, learn a whole new area like databases, rewrite the whole kernel, etc.).
5. You need to know why you want a Ph.D. You need to have vision and ideas and you need to be able to express yourself.
6. Obviously, many people are still unsure straight after a B.A.. I was one of them, so I understand. For such people working in a research lab or in an industrial lab which involves doing research for a few years will help them decide. If you are unsure, I highly recommend working for a few years before starting a Ph.D.. Do not apply to graduate school until you are sure you know what you want.

Prof. Mor Harchol-Balter's own story:
After I finished my B.A. in CS and Math, I went to work at the Advanced Machine Intelligence Lab at GTE in Massachusetts. At first I was very excited by my paycheck and the great feeling of being independent. I also really enjoyed my area of research at the time: pattern recognition and classification. I was working with frame-of-reference transformations involving eigenvectors of autocorrelation matrices. It was exciting! However I quickly realized that I wanted to know more. I wanted to know why some algorithms produced good results and others didn’t. I wanted to come up with my own algorithms. I worried that I didn’t have enough of a mathematics background to answer my own questions. In summary, I wanted to delve deeper. Everyone around me thought I was odd for wanting these things. I left after 2 years and went to graduate school. That first month of graduate school I looked around and realized that everyone there was just as weird and obsessed as I was, and I knew I had made the right decision.

Specialized mailing lists

If you made your mind what field you want to work within, consider joining specialized scientific mailing lists related to your field. It's very useful to be on the same mailing list with pioneers of the field. Subscription at such mailing lists give you the following benefits:
- Notifications about call-for-papers/participation at conferences as well as deadline extensions.
- Notifications about special issues at journals.
- Announcements on PhD/MSc opportunities and scholarships.
- Announcements on Research-Assistantship, Post-Doc and relevant job vacancies.
- Resources and tools made available for research community.
- Discussions on research directions by professionals in the field.

Following are examples of fields and respective mailing lists:
- Natural Language Processing: Corpora-list
- Data Mining/Databases: KD-Nuggets, DB-World
- Social Network Analysis: Socnet
- Machine Learning: ML-news, UAI
- Neural Nets: Connectionists
- Information Retrieval: SIG-IRList

Thanks to Hossam Sharara for inspiring this post.

Previous research experience (if you plan to apply for a PhD program)

This post is the third episode featuring Prof. Mor Harchol-Balter's talk advising people applying to PhD programs in computer science or related areas.
[first episode - second episode - full article]

As I’ve said earlier, to get into a top graduate school you need prior research experience. This is not necessarily true for schools below the top 10, or maybe even the top 5. Note that prior research experience does not mean that you need to have published a paper. It does not even mean that your research needs to have yielded a result – results can sometimes take years. We just need to have confidence that you know what doing research is like. At CMU we receive hundreds of applications each year from 4.0 GPA students who have never done research. These are all put into the high risk pile and are subsequently rejected.

So the question is, where can you get this research experience?
There are five places where you might get research experience:

1. As an undergraduate, you can do research with a professor. I did this. You can even get course credit for this, and sometimes if you’re really lucky you can get paid a little (e.g., during the summer).
2. As an undergraduate, you can apply for a summer internship at a research lab, e.g., AT&T. I did this. They will pay you a little and you will learn a lot about doing research. This was a great experience for me! Here’s the web site for the AT&T summer program that I attended: http://www.research.att.com/academic/ When you go to this web site, click on “Special Programs and Fellowships.”
3. After graduating, you can get a job, where sometimes you can do research on the job. I did this.
4. As an MS student, you will work on an MS project.
5. You can work alone or with a friend. Ask professors in your classes to tell you about interesting open problems and new research (most professors enjoy doing this). Ask them to tell you names of conference proceedings. For example in my area (performance modeling of computer systems) a relevant conference proceeding is Sigmetrics. Sit down and start reading these proceedings. You will come across all sorts of interesting problems. Think about how you can improve upon the solution proposed in the paper.

...
Warning for international applicants: The admissions committee needs to be able to evaluate your research. If your publications appear in conferences/journals which we are not familiar with and have no access to, then we cannot evaluate the quality of your work. In my experience, this usually leads us to discount such publications. If you don’t want this to happen, here are two things you should do:

• Publish in internationally recognized conferences – ask your advisor.
• Create a web site that has links to all of your papers in English in either postscript or pdf. Explain in your application that all of your papers can be found on your web site.

How to become a star grad student

In this article, Cal Newport tries to answer this question. He takes James McLurkin as a case study, and starts by explaining how far one could be recognized in his field, and why this is so important.

“Four years earlier, Time Magazine profiled James as part of their Innovators series. The next year, he was featured on an episode of Nova ScienceNOW. Earlier this year, TheGrio, a popular African American-focused news portal, named James one of their 100 History Makers in the Making.”

“In other words, James is famous in his field. So it’s not surprising that in 2009 he landed a professorship at Rice University — one of the country’s top engineering schools — in one of the worst academic job market in decades.”

James' stardom started when he designed a swarm of microbots 'Ants' for his senior thesis project (i.e. graduation project). The devices were designed to perform complex behavior using simple rules. The paper in which he documented this work spread out to the public media, making a star out of James.

“I went to the lab as an undergrad to interview for a position,” James recalls. “Anita Flynn told me they’re not hiring. So I came back with some robots I had built, and some I was halfway through building, and she said, ‘okay, you can work in the lab, and use our parts, but we can’t pay you.’”

Once in the lab, he worked real-hard on one project after another; each stretching his abilities a little bit. He wasn't alone in the lab though. Anita Flynn was shrinking the size of electronic motors, enabling the micro-robot revolution, while Maja Mataric was a leading thinker on robotic swarms.

“By the time he conceived of the Ants project for his thesis, James was an accomplished robot engineer with a number of successful projects under his belt. He also had a cutting-edge knowledge of microrobotics, and was “marinating” in a lab environment obsessed with biologically-inspired systems. With this in mind, the idea of building a robot swarm that behaves like insects was not a big hairy audacious goal to him.”

It was James' knowledge and expertise in cutting-edge techniques in his field that enabled him to take an unprecedented step. To him (and others with equivalent level of knowledge and expertise), Ants was an obvious incremental step. To the rest of the world, including less-aware people in robotics field, it was a huge breakthrough. Cal concludes that, to become a star, you should focus on getting to the bleeding edge of your field as quickly as possible.

“Many graduate students, for example, never arrive at the bleeding edge of their field. Instead, they reach a comfortable level of knowledge — enough to understand relevant research, and make their own acceptably-complex contributions, but not enough to make bold advances. Thousands of chemists could understand Watson and Crick’s 1953 paper on the double helix, but only a handful had the knowledge needed to have discovered it for themselves.”

Now, the obvious question is how do you get to the bleeding edge?

“Every semester, my supervisor, Anita, had me write out goals,” James told Cal. “We would go back at the end of the semester and look at what I did and didn’t do. She would tell me, ‘it’s fine that you didn’t get this all done, but what’s not fine is your inability to estimate how long something will take.’”

James deliberately chose projects that were hard enough to stretch his ability, but reasonable enough to complete in the available timeframe.

“With this in mind, I argue that the secret to James McLurkin’s success is his ability to choose the right projects. By resisting work that reinforced what he’s comfortable with, yet also sidestepping overly-ambitious projects, he consistently advanced his skill until he arrived at the bleeding edge of research robotics. Once there, the “breakthrough” projects that cemented his reputation became obvious next steps. Stretch projects are an effective way to integrate deliberate practice into fields without clear competitive structures and coaching”

To emphasize, Cal gives two definitions:
Stretch Project: A project that requires a skill you don’t have.
Stretch Churn: Number of stretch projects you complete per unit time.

In order to make it to the bleeding edge, you need to maximize your stretch churn. You need to be in a continuous discomfort learning new things and resist the tendency to reinforce what you already know.

Uh.. I think this comment (in response to Cal's article) is also worth quoting
“Nianu: What kind of stretch projects would you recommend I start to tackle? I have a hard time thinking of what would be a good way to start as I am still early in my college career.”
“Cal: College and graduate-level courses are stretch projects in themselves. They force you to acquire new skills, but everyone completes them within a relatively short time frame.
Attack your courses with the mindset. Savor the hard focus required to master the material (coupled, of course, with smart study tactics to eliminate wasted time and effort), knowing that you're building the skills needed to move toward the bleeding edge.”

Recommendation letters

This post is the second episode featuring Prof. Mor Harchol-Balter's talk advising people applying to PhD programs in computer science or related areas.
[first episode - full article]

"Perhaps the most important part of your application is the letters of recommendation. You will need 3 letters of recommendation for the Ph.D. program, and typically 4 letters of recommendation for a fellowship.

Whom to ask for a letter
Ideally you would like to make all your letters of recommendation count. Consider the following two letters:

• Letter 1: “I highly recommend student X for your graduate program. Student X received an A+ in my undergraduate algorithms class. He was ranked Number 2 out of 100 students. He got the highest score on the final. He worked very hard all semester, never missed a class, and was always able to answer the questions that I asked in class. This conscientious attitude makes him an excellent candidate for any graduate program. ”
• Letter 2: “I highly recommend student Y for your graduate program. Student Y received a B in my undergraduate algorithms class. He was ranked Number 29 out of 100 students. Halfway through the semester we started working on network flows. Student Y seemed extremely excited by this topic. He disappeared for 4 weeks and even missed an exam. However when he came back, he showed me some work he had been doing on a new network flow algorithm for high-degree graphs. He had done some simulations and had some proofs. I’ve been working with student Y for the past couple months since then and he is full of ideas for new algorithms. I think student Y’s initiative makes him an excellent candidate for any graduate program.”

Which letter do you think is stronger? It turns out that Letter 2 is very strong. Letter 1 actually counts as 0. At CMU we mark all letters like letter 1 with the acronym D.W.I.C.. This stands for “Did Well In Class” which counts for 0, since we already know from the student’s transcript that he did well in class. By contrast, student Y’s letter gives us a lot of information. It explains that the reason student Y didn’t do better in class was that he was busy doing research. It also tells us that student Y started doing research on his own initiative, and that he is quite good at doing research. The professor was impressed enough with student Y’s ideas that he took him on as a student researcher despite student Y not having high grades. You want your letters to all be of type 2 (this doesn’t mean that you should skip class!). Remember that letters of type 1 will not count. You want words like 'self-motivated', 'strong research potential', 'own initiative', 'independent', and 'driven' to appear in your letters. These are the words that we circle when reading recommendation letters. You therefore want to ask letters from people who have seen you do research. These may be professors or employers.

One caveat: It makes some difference whom you ask for a letter. As a general rule (there are always exceptions due to people’s fame), letters from professors count the most. Next highest are letters from research scientists. After that come letters from lecturers, systems scientists, employers, or postdocs. Please do not get a letter from a graduate student. If you found yourself doing research where you were supervised by a graduate student or postdoc, you should ask the professor for whom they work if she can co-write the letter. The reason is simple: professors are the ones reading the letter, and they are most likely to know other professors.

There is an issue for students who have been working for a while. You will certainly want a letter from your employer, but you will also want two letters from professors. This was an issue for me when I applied to graduate school. What I did was to keep touch with a few professors during my time at work. When I was ready to apply to graduate school, I contacted the professors who knew me well and scheduled a meeting with them to discuss the research that I had done while I worked. I gave them each an oral presentation. I also gave them each writeups of each of my projects.

How to ask for a letter
Asking for a letter of recommendation won’t be a problem if you have been doing research with this person, but that won’t be possible in every case. Here’s a guideline which will maximize the contents of your letter. This works on the theory that professors have very little time and little memory (both of which are good assumptions):

1. Prepare a packet for each recommender. This packet should contain all the relevant information about you that could help the recommender. Be careful not to make the packet too large. Here’s what should be in it:
   • Your statement of purpose.
   • A summary of every research project you worked on and with whom, regardless of whether this was at a school or research lab. If you have published a paper, or have a technical report, please include that too.
   • A sheet of paper listing all math/cs/engineering/science classes you have taken with the names of professors and grades.
   • A list of extracurricular activities and awards/competitions.
   
   At the top of the packet should be:
   • A recent photo of you – professors receive many such packets and don’t remember you the second after you leave the office.
   • Directions. E.g., please seal and sign and send to this address by Jan. 5. Put an earlier date than the real deadline – professors are notoriously late.
   • Confirmation information: Please send me email at blank address after you send this off. If I don’t hear from you by Jan. 5th, I will send you an email reminder. (You need this confirmation information because otherwise you’ll never know if the recommendation was sent and you’ll be sitting around biting your nails wondering.)
2. Go to your potential recommender with your packet and ask him/her the following question: “Do you feel comfortable writing a strong letter of recommendation for me to graduate school?” You need to phrase the question this way so that the potential recommender has a way out. Do not be upset if the potential recommender says no. It is good that he/she let you know. This is much better than getting a weak letter.
3. Check with the school to confirm that they have received a letter from each of your recommenders.
4. Remember to at least send your recommender a thank you card! It’s a lot of work to write a decent recommendation letter, and you may need more letters in the future."

MSc then PhD?

Typical situation: Hamdy was a top ranked student in his university, in a developing country, and now he'd like to get a PhD from a good university in US/Canada (with financial aids). He has two options:

1. Apply for a PhD.


2. Apply for an MSc. And after getting the MSc degree, he applies for a PhD.

Why would Hamdy go for option 2:

Avoid frustration
MSc is less of a commitment than PhD. It's easier to work on, and takes much less time to get done. However, it gives a clear idea on how Hamdy's life is gonna be as a PhD student. Given that 50% of PhD students in some fields drop out before completing their degrees, this becomes a significant consideration. Hamdy prefers to go for the less risky degree (ie. MSc) and then decides whether he should continue for a PhD or get a good job in the industry with his MSc degree, from a good university in US/Canada. I know one person who left PhD for the industry, and many PhD students who consider doing the same.

Chances of acceptance
MSc is less of a commitment than PhD, this holds true even for the admission committee of the university he applies for. Several persons I know applied for a PhD and were offered only an MSc. So, applying for an MSc would increase Hamdy's chances of getting admitted for graudate education.

Improve credentials
What if Hamdy wants to get his PhD from the reputable university A, while his credentials qualify him only for the less-reputable university B. He can first apply for an MSc with B and get the MSc degree to improve his credentials, then apply for a PhD in A. He can then transfer his MSc credits from B to A (this is possible for many A-B universities pairs in US/Canada). Note that -if this is what Hamdy plans to do- he shouldn't apply for a PhD in the less reputable university (B), because his supervisor will be grooming and expecting him to work for him for ~5 years, and then Hamdy will let him down and -moreover- ask him for a recommendation letter to the more reputable university (Hamdy can't go to A without a recommendation letter from his supervisor at B). I'm not sure what the supervisor will write in the recommendation letter then about Hamdy's commitment.

Freedom
Younos Aboulnaga was the first to introduce me to the concept of "taking a break while studying". In his undergraduate study in Egypt, Younus paused before last year, spent a year working in Brazil, and then went back to earn his BSc degree. Almost all PhD students get fed up after a few years of working for it. They wish they could "take a break" just like Younus did. However, they can't just apply for any job and spend a year or two, then come back to continue the PhD. They have to find an internship inline with their research, competing with many PhD students for a limited number of internships, probably with low pays. If Hamdy goes for option 2, he could pause for as long as he wishes, then gets back to continue his graduate education when he feels he's ready to continue.


Why would Hamdy refrain from option 2:

Time
Option 2 usually takes more time, even though he will probably take the MSc in option 1 as well.

Commitment to research
Most of the rationales for taking option 2 assumes it's likely for the student to drop out. If Hamdy knows somehow he's comfortable with doing research and will withstand the pain it takes to get a PhD, these rationales weight nothing for him.

Over qualified
Some universities (e.g. UCSB) won't admit you for a masters program if you already have had a masters degree in the same disciple from another university.

Request: If you know something that contradicts with what I'm stating here, or complements it somehow, please do comment on the post or drop me an email at ammar DOT w AT acm DOT org, and I will be happy to update the original post reflecting your valuable point of view.

How to enhance runtime performance?

Since we do computer science research, we typically implement our explorations and research advances as computer programs. Many research fields have to deal with an immense amount of data (e.g. information retrieval, natural language processing, data mining), adding scalability as a major requirement when you develop a prototype/an exploration.

This post is all about how to enhance runtime performance of your program. It summarizes a discussion I had with some colleagues in this regard:

1. "Premature optimization is the root of all evil (or at least most of it) in programming", says Donald Knuth. Instead, use a profiler to identify which pieces of the code need your attention.

2. Optimize the algorithm first, not the code. Sometimes we think a problem is too trivial to analyze its algorithm, but it makes much sense to reconsider the 'trivial' algorithm when it runs millions of times. For example, this code

double metric1 = calculateMetric1(input);
double metric2 = calculateMetric2(input);
if(metric1 > threshold1 && metric2 > threshod2)
  return true;
else
  return false;

Assuming 'calculateMetric2' is an expensive procedure, refactoring the code as follows will significantly enhance performance:

double metric1 = calculateMetric1(input);
if(metric1 <= threshold1)
  return false;
double metric2 = calculateMetric2(input);
if(metric2 <= threshold2)
  return false;
else
  return true;

3. Concentrate on high level optimizations. Most popular compilers do a lot of optimization on your behalf. For example, Microsoft's Visual C compiler will replace the following code with "x = 27000000":

x = 0;
for(int i=0;i<300;i++)
  for(int j=0;j<300;j++)
    for(int k=0;k<300;k++)
      x++;

4. Consider changing the data structure that hosts your data (e.g. a hashmap instead of an array).

5. Caching. If there’s an expensive operation with inputs that are likely to repeat, it might be worth caching. Note that a cache with low hit-ratio might degrade overall performance.

6. Distribute the load on the multiple cores of your processor, or on multiple machines.

7. Get rid of redundant calls. For example, when the code aggressively uses case-insensitive string comparisons, lowercase your strings only once.

8. Sometimes, memory allocation is very demanding. The solution in these cases is usually to create your own memory allocator or use memory pools for certain objects.

9. Sometimes, initializing/resetting arrays is very costly. It’s not always necessary to initialize/reset an array before using/reusing it. For example, to compute the Levenshtein distance between two words, there’s a dynamic programming algorithm which uses a 2D array. Even though the same array is used for different word pairs, no need to zero the array since it’s built in a bottom-up fashion.

I'd like to acknowledge my friends/colleagues (Ahmed El Deeb, Ahmed Sabry & Diaa Samy) for the constructive discussion we had around this topic.