Prof. Philip Guo

Video Link: https://pg.ucsd.edu/advice-for-first-year-PhD-students.htms

1-D model

Keep moving. This is probably not a good idea. Direction is fixed.

N-D model

Move at a good pace, stop, and look around. Pointing to a correct direction is important. One of the only way to develop intuition is starting doing things around it or in it.

Advisor meetings

Talk about a concrete thing.

  • notes about experiments
  • theory/paper draft
  • software prototype
  • graphs, figures

Twelve sticky notes

  • Start simple
  • Look at raw data
  • Be concrete
  • Incremental results
  • Find something interesting
  • Push hard on clean signals
  • Actionable
  • Comparative advantage is key
  • Find solid levers to pull
  • Document successes and failures (YG: Personally, I learned this from my labmates.)
  • Inspect outliers
  • Eliminate sources of bias

Advice for early-stage Ph.D. students

This part is completely copied from a post of Prof. Philip Guo. I found this advice is useful, so I maintain a copy here in case of deletion of the original post. All copyrights belong to Prof. Philip Guo. Thanks!

I know this sounds presumptuous, but if you just started a Ph.D. program, especially in science or engineering, bookmark this page and read it once a week. You won’t internalize much of the contents at the outset, but parts will start resonating with you as you progress through grad school.

[Warning: these notes are somewhat messy and disorganized.]

UNDERGRAD VERSUS PH.D. RESEARCH

If you’re now in a Ph.D. program, you’ve likely had positive experiences with research as an undergraduate. Also, you’ve been a good student in school, scoring at the top of your class on exams and projects. And you’ve probably been praised throughout childhood for being a smart kid.

My first warning for all new Ph.D. students is that Ph.D.-level research will be much more difficult than undergraduate research both in terms of the technical aspects and also the emotional stresses. So do not underestimate the inevitable hardships that await you in the coming years. One common cause of failure amongst early-stage Ph.D. students is underestimating the difficulty of the process, especially the early stages. At this point, you won’t fully grasp the nuances of why this process is so hard, so the only advice I have is just to prepare for immense challenges in the years ahead.

Everyone starts a Ph.D. program with an incredible amount of potential. But most students never come close to approaching their full potential for creative research achievements. And the reason is never because they’re not smart or technically capable enough; everyone who gets admitted is smart enough. Rather, factors such as lack of resilience, perseverance, metacognition, and self-discipline are the main contributors to failure at the Ph.D. level. Fortunately, these traits can be fostered via self-reflection and mentorship.

UNCERTAINTY AND ISOLATION

All researchers-in-training must constantly grapple with:

  • uncertainty – You have no idea whether the hard work you are putting into your project even matters.

  • isolation – Nobody around you understands or empathizes with what you are doing, since they either lack technical context or are too busy with their own creative struggles.

If you can properly manage these two emotions and make consistent forward progress every single day, get private feedback from a mentor every week or two, and get external feedback from paper submissions a few times per year, then you can successfully finish your Ph.D.

The bad news is that it’s impossible to fully eliminate uncertainty and isolation when doing research. But if it’s any consolation, recognize that these feelings are completely normal; all of your fellow grad students are facing them as well.

(Update on 2014-02-15: A third daunting problem that many early-stage researchers face is that of project scoping. Research is often open-ended, but concrete deliverables must be produced. So how much work is required for an acceptable prototype or experiment? How much is enough for a respectable paper submission? How much is needed for a master’s thesis? For a Ph.D. dissertation proposal? For a completed dissertation? In contrast, K-12 and university classes are all well-scoped by instructors’ expectations.)

DEVELOP RESEARCH TASTE

When you first start your Ph.D., you might want to immediately dive into implementing your own creative ideas. The problem is that your taste isn’t yet calibrated to what is considered “good research” by your academic community. Even if you think your taste is impeccable, that doesn’t matter one bit; to publish papers and earn a Ph.D., you need to do work that resonates with senior researchers in your field.

But wait … wasn’t the appeal of being a researcher that you can do creative work rather than doing what your superiors order you to do?!? Well, sorta. To innovate in any creative field, you must first understand the tastes of the establishment, and only then can you inject your personal flair. (Matt Might illustrates nicely.)

So how do you develop research taste?

Read – Ask your advisor for a set of well-respected papers in your field published within the past few years. To earn a Ph.D., you will need to write papers that look like those, so learn their methodologies, technical conventions, and presentation styles now. Note that old papers might be fun to read, but they’re less helpful for honing your taste, since you won’t be able to publish papers like those anymore; their styles are often out of fashion.

Assist – Assist your advisor, senior students, and postdocs in your lab on their projects. Make yourself as useful of a helper as possible without worrying about taking creative control. The best-case outcome here is that you end up as a non-lead coauthor on their papers and learn a lot about research methodology and conventions. Assisting also eliminates the uncertainty and isolation that often paralyze early-stage students.

Grind – Even when you start developing good research taste, the early work you produce won’t be good. That’s okay! As cliched as this sounds, you need to grind hard for years before getting good at anything worthwhile. My favorite Ira Glass quote beautifully captures this idea:

What nobody tells people who are beginners – and I really wish someone had told this to me ... is that all of us who do creative work, we get into it because we have good taste. But there is this gap. For the first couple years you make stuff, and it's just not that good. It's trying to be good, it has potential, but it's not.
But your taste, the thing that got you into the game, is still killer. And your taste is why your work disappoints you. A lot of people never get past this phase. They quit. Most people I know who do interesting, creative work went through years of this. We know our work doesn't have this special thing that we want it to have. We all go through this. And if you are just starting out or you are still in this phase, you gotta know it's normal and the most important thing you can do is do a lot of work. Put yourself on a deadline so that every week you will finish one story.

It is only by going through a volume of work that you will close that gap, and your work will be as good as your ambitions. And I took longer to figure out how to do this than anyone I've ever met. It's gonna take awhile. It's normal to take awhile. You've just gotta fight your way through.

Also, read my article Lessons from the Grind about why grinding is a precursor for creativity. In short, without a ton of hard work, you can’t even begin to generate original ideas that go beyond the state-of-the-art in your field.

Finally, the sci-fi author William Gibson has a great mini-essay on developing creative taste as a fiction writer. Here is the concluding paragraph, with me substituting “writing” for “research”:

And therein, I think, lies most of how one learns to do research. We have to learn to do research, but we have already, to varying degrees, had to learn to read research. And I felt like quite a good reader of research, when I began to do research, or at least a good reader of that research which I most keenly enjoyed. And thus are we shaped as researchers, I believe, not so much by who our favorite researchers are as by our general experience of research. Learning to do research, we learn to listen for our own acquired sense of what feels right, based on the totality of the pleasure (or its lack) that research has provided us. Not direct emulation, but rather a matter of a personal micro-culture.

UNDERSTAND YOUR ADVISOR

Behind that facade of authority, your Ph.D. advisor is a human being with their own needs, wants, and biases. The better you understand your advisor, the smoother your Ph.D. experience will be.

Perhaps the most important question is: Does your advisor have tenure? If not, then their top priority is earning tenure so that they can keep their job. If you work with an untenured advisor, then your work will directly contribute to their tenure case. Thus, untenured advisors are usually more hands-on and provide more structure for their Ph.D. students’ careers. Also, untenured advisors are easier to read, since their main professional goal is to build a compelling research portfolio for their tenure case.

If your advisor already has tenure, then they might have varying motivations. Do they want to build up a larger and more ambitious research program? Are they trying to break into a new line of research? Are they now more focused on professional service, advocacy, or teaching? Tenured advisors tend to be a bit more hands-off since their careers aren’t as dependent on their Ph.D. students’ performance. But of course, there are many tenured advisors who are still just as hungry for success as they were in their younger, untenured years. It’s ultimately up to you to gauge their motivations and priorities.

BE REALLY PATIENT

For a six-year Ph.D. (which is typical for science and engineering programs in the U.S.), what you do in the first three years probably won’t count toward your dissertation. Seriously, pretty much every Ph.D. graduate I’ve spoken with shares this same experience. So be patient.

If you’ve finished three years and still don’t have a dissertation project with published papers under your belt, it’s okay. As long as you’ve been learning to develop good research taste by reading, assisting other people’s projects, and trying (and probably failing) to push forward your own projects, then you’ve set yourself up well for the second half of grad school.

For instance, I started my first successful Ph.D. project at the beginning of my fourth year and didn’t get the paper published until the middle of my fifth year.

You will inevitably encounter peers who are “ahead of you” in their Ph.D., publishing more papers and finding success earlier on in grad school. That’s okay! It’s not a head-to-head competition; there is no class curve. So be patient and march forward every day, one bit at a time.

(Note that if you work with an untenured advisor, then they have a strong vested interest in getting you up to speed and productive as quickly as possible, since they need more publications to build up their tenure case.)

MAKE PROFESSORS WANT TO HELP YOU

You can’t get good as a researcher without help from professors, but the conundrum is that professors like helping students who are already good at research.

If a professor had the choice between spending an hour with an all-star student who is about to submit a strong paper and a naive early-stage student, which meeting would they look forward to? Which meeting would be more fun for them? In general, professors look forward to helping students who, paradoxically, need less help.

Here’s an open secret: Professors are neither hired nor promoted based on how well they mentor grad students. Advising quality only matters to the extent that good advising can produce papers, but I’ve seen plenty of bad advisors successfully crack the whip to churn out papers as well. When I interviewed for faculty jobs, never once did my interviewers ask how I would advise Ph.D. students, or about fostering grad student health in general. In fact, the topic of grad students never came up, except when they were complaining about bad ones. Sadly, those are the students who need the most help but don’t know how to get it.

So what’s the lesson here? You need to make professors want to help you. Repeat: You need to make professors want to help you.

How? One way is by demonstrating that you have impeccable work ethic and great potential for future success, so that they feel like their time is being well spent. Another way is to discover what truly excites them and adapt your interests to theirs. Read my article Lead From Below for more details on this technique.

If can’t make professors want to help you, then they would rather devote their energies to their other students. Heck, even as they’re meeting with you, they might be wishing that they were instead hanging out with their all-star student.

(This advice applies to all jobs, not just research. The most successful employees are often those who make their bosses or mentors want to help them.)

FIND PEER SUPPORT

The happiest and most successful Ph.D. students are those who have maintained a strong peer support group throughout grad school. Remember, isolation comes by default, so you need to proactively seek out peers for camaraderie. Your department and advisors cannot do much to help, despite their most sincere efforts: Lab lunches and snack breaks are superficial patches and don’t do much to eliminate the endemic feelings of isolation. So seek strength from your peers, not from your superiors.

UNDERSTAND YOUR JOB

The happiest and most successful Ph.D. students understand that this is a job, albeit a unique one with different expectations than typical industry jobs. The most angst-ridden students still think of the Ph.D. experience as an extended form of school and a shelter from getting a “real job.” Well, this is a job! I’ve noticed that students with a few years of industry experience generally have a better time in grad school than those who came straight from college. This is a gross over-generalization, though. Counterexamples abound on both sides.

In most science and engineering fields, students are funded by their advisors’ grants, which stipulate specific projects or research directions that they must work on. As a student, your funding source, advisor’s expectations, and the current tastes of the research community all dictate what kind of work you can potentially do. You don’t have total freedom; but then again, nobody does. Once you’ve internalized your role at this job, then you can figure out ways to be creative within those constraints.

If you don’t have any industry experience, I highly recommend spending a summer or two interning at a company during grad school, especially in a non-research role. Internships not only help you understand how projects get done with much shorter time horizons than research projects, but also help you develop skills that are useful outside of academia.

MAKE YOURSELF ACCOUNTABLE

Here’s one major difference between industry and academia:

  • In industry, you are usually working closely with a team and given concrete tasks with weekly and sometimes even daily deliverables. You know exactly what you need to do when you get to work every day.

  • In academia, you are usually working alone and do not have concrete day-to-day deliverables that you must complete. Of course, your advisor or senior colleagues might be relying on you to get stuff done, but the pressure isn’t usually on a daily basis.

Thus, it’s much easier to slip through the cracks in academia because you are often not on anybody’s critical path. (By critical path I mean the path of work that is critical for their career advancement or fulfillment at the given moment in time.) One notable exception is if you have a untenured advisor who needs your project as part of their tenure case; then by definition you are on their critical path to earning tenure.

Also, research is inherently less concrete than industry work, so it’s harder to track daily progress. If you’re not performing up to par in industry, your teammates notice immediately and will figure out some way to get you back on track … or get you fired. After all, their careers are depending on you, so they can’t afford to have you drag along as dead weight. In contrast, as a grad student, if you’re in a slump for a few weeks or even a few months, your advisor might not notice if they are busy with other aspects of their job.

One hack is to find ways to make yourself accountable to other people, thereby simulating this desirable aspect of industry jobs. If you can tie your success directly to someone else’s, then they will be more likely to keep you on track and making consistent progress toward your mutual goals. Again, working with an untenured advisor automatically makes you accountable, since their tenure case depends on your work.

KEEP MOVING

If you’re actively working on a project (i.e., not on vacation) and spend more than a few days not doing anything concrete related to it, then you’re stuck. It’s critical that you talk to your advisor or another mentor immediately so that they can help you get unstuck ASAP. There’s no shame in being stuck; it happens to everyone. It’s your advisor’s job to get you unstuck.

Joel Spolsky’s Fire And Motion eloquently sums up the philosophy of keep moving:

In infantry battles, [the general] told us, there is only one strategy: Fire and Motion. You move towards the enemy while firing your weapon. The firing forces him to keep his head down so he can't fire at you. (That's what the soldiers mean when they shout "cover me." It means, "fire at our enemy so he has to duck and can't fire at me while I run across this street, here." It works.) The motion allows you to conquer territory and get closer to your enemy, where your shots are much more likely to hit their target. If you're not moving, the enemy gets to decide what happens, which is not a good thing. If you're not firing, the enemy will fire at you, pinning you down.

And Michael Nielsen in Principles of Effective Research:

In my opinion, there is little that is more important in research than building forward momentum. Being clear about some goal, even if that goal is the wrong goal, or the clarity is illusory, is tremendously powerful. For the most part, it’s better to be doing something, rather than nothing, provided, of course, that you set time aside frequently for reflection and reconsideration of your goals. Much of the time in research is spent in a fog, and taking the time to set clear goals can really help lift the fog.

Many Ph.D. students fail not because they’re not smart or hardworking, but because they get stuck for extended periods of time and then grow demoralized. Keep moving.

See this talk for more details: Advice for first-year Ph.D. students.

COMMONLY OBSERVED STRUGGLES

[Written in June 2020 after finishing my sixth year as an assistant professor. By this point I’ve participated in six end-of-year meetings where faculty collectively discuss the progress of all Ph.D. students in our department. Here are some of the most common struggles we’ve seen.]

  • Lack of focus: Jumping between projects too frequently, quitting a path too early and switching to whatever else looks new and exciting, juggling multiple unrelated projects, trying to emulate what they see some faculty doing by collaborating as a co-author on a bunch of unrelated papers rather than working toward one’s own first-author papers.
  • Not telling advisor that they’re stuck: Some students are stuck for long periods of time without telling their advisor, perhaps due to fear or shame. The more time that passes, the harder it is to talk to their advisor about this, and the more they resort to procrastination (see next point), which further stalls their progress.
  • Using other activities as procrastination: Research is inherently hard, so it’s all too easy to procrastinate on it by spending a ton of time on other activities that have shorter-term and more tangible rewards. Of course, it’s great to have extracurricular activities; that’s what makes life worth living! Believe me, I’m not trying to discourage that at all. But if you find yourself spending too much energy on those and not enough on research, then that could be a sign that your research progress has stalled.
  • Not having a single lead advisor: Especially for students working on interdisciplinary projects or co-advised by several faculty (maybe across multiple departments), it’s easy to fall through the cracks since no single person takes the lead advisor role. If no faculty can clearly summarize what you’re currently working on, then that’s a bad sign. Ideally you would be on one single advisor’s critical path so that they’re highly motivated to track your research progress. (By critical path I mean the path of work that is critical for their career advancement or fulfillment at the given moment in time.)