About my Academia Life

I received many inquiries about my academic life. I think it might be great to share some details behind my resume. What happened before my PhD? Why PhD and why not PhD? How’s my PhD? Is it easy? I want to write down some thoughts…

Tough Time Before Ph.D.

I was never considered a “good” student. I was always considered a “unique” student. Though I don’t feel I am unique at all, I know I have limited talents in answering multiple choice questions. I used to force to be a piano prodigy or national-level swimmer. But later on I was not competent but gladly at least I learned how to play piano and swim.

From the parent-teacher meetings, my parents were told that I was “too talkative”, “too active”, “too innovative”, “had too many questions.” My parents usually felt sad hearing this (Honestly speaking, I think it’s mixed but leans towards positive). Later on, they gave up going to the parent-teacher meetings as they already knew what the teachers would say. Then I transitioned to an international school, and it was more like a home school style where I usually spent time having fun alone. I became like an independent artist, spending time mainly on films, drawings, and music. I recorded my first album from 2016-18. It probably sleeps somewhere on my shelf.

In 2017, I started my undergrad in the US. Initially in my college application, my mom suggested biomedical statistics as my major. After I started my first semester, I truly believed Computer Science was the future. The first Computer Science class was super fun, playing around with Scratch. I enjoyed the class and later on in my senior year I became a head TA and part-time lecturer. That’s the starting point of my academic “career”.

I still remember in my second year, I failed Calculus II (getting a D). If I retook it again and still got below C, then I would be out of the Computer Science major. That was summer 2019, extremely hot in New Jersey. I felt super anxious and was afraid of being forced to drop out of the major. Fortunately, in the following Fall semester, I passed.

At the end of my sophomore year, I had a friend randomly come to my dorm to have barbeque together. She asked about my plan, and I told her I only knew I wanted to graduate in 3 years and maybe apply for a PhD. She looked me in the eyes and said, “you have limited experience in conducting research and clearly you are not prepared for it. You need to find a research advisor and do research.” That changed my life. I started to contact several professors and fortunately Prof. Mubbasir Kapadia accepted me as an undergrad research assistant. From then on, I officially started my research career and submitted to multiple places such as AAAI undergrad consortium (rejected), Rutgers Aresty Research Center (got $1000), and Interdisciplinary Contest in Modeling (ICM) (meritorious winner).

Then, I started my master’s during Covid. I think online classes are great. I tried my best to do research and write a paper, but I didn’t know how to start. I wasn’t sure about my research interests and career plan. Many times, I felt hopeless. I also applied to several big tech companies to do internships but none of them succeeded.

Reflecting back, my undergrad and master’s time were not really good. Comparing to many people who started their research and published a lot, I “wasted” my time. I spent most of the time learning and studying. But I think starting research early is not really helpful, as the whole mindset becomes about publishing. But that’s not the whole point of research. Research is about exploring curiosity, doing something fun and impactful, and finding novel solutions. It’s supposed to fail a lot. Research means wasting time on replicating other people’s failures, reading other people’s fancy papers, and even deep-thinking. (I am not defending why I am not productive; I am defending that when I am not writing papers and looking productive, I am still doing research)

OK Time During Ph.D.

After starting my Ph.D., my life became different. I started to know how to write an HCI paper and learned more about the publishing mechanisms. Just judging from the research question, I could tell how good a project would turn out. I also started to know what kind of mentorship I needed. Before my Ph.D., I didn’t even know what kind of assistance I needed to write a paper.

I am super fortunate that I received a lot of support during my Ph.D. There were many things that I didn’t receive support for and later on looking back, I felt extremely grateful that I failed. But I tried 120% before I admitted I failed. At that time I felt I was a loser. I switched advisors, I tried to switch majors but failed, I suffered from no GPUs, I encountered a no research funding situation… These failures were invaluable and taught me to learn quickly and avoid these situations in the future.

In the third year of my Ph.D., I was only eligible for TA funding, which required me to TA 2 classes per semester. It became very hard for me to do research and collaborate with clinics and industry partners. Oh my! Fortunately I got to know a IvyPlus Exchange Ph.D. scholarship but I had only 12 days to apply for it, including contacting the host institutions’ professors to host me. I remember that morning I burst out in tears because I felt hopeless and helpless. But later on, Prof. Marzyeh Ghassemi hosted me and provided me funding for my research. Currently, the funding situation is hard and only trying and bold movements may at least make a difference. I think it’s because the previous failures taught me how to recover quickly and not think change is a bad idea. It means something new, something even better, or something that must happen in destiny.

I am very execution-focused. If I decide to do it, then it will be now. I always tell myself, if you decide to do it, then the best time is now, and the best way to do it is all-in. You shouldn’t wait and see, because usually afterward the result is giving up. You shouldn’t put in partial energy and be ready to leave; the best strategy is all-in and trust the vision. When I have this kind of commitment, nothing really fails. Failure is not an option, because I will never give up until I win. I think a PhD may need a high requirement on execution; you have to go experiment, go try your hypothesis, go test your methods, go admit it doesn’t work but tomorrow I will be at the office at 9 am and start experimenting again. But the dark side of being highly execution-focused is I have no patience for waiting. I used to work with someone who told me next week I could ping again to check the results and never got a response anymore. If it’s not tomorrow then I am going to do it myself.

Is Ph.D. Necessary?

I have 2 versions of answers, as always. The good version is “It’s necessary because I want to do research”; the bad version is “I can’t find a job after graduating from my bachelor’s/master’s.”

But here, let’s focus on the good version. First of all, I think research is fun. I enjoy doing applied research. But research is not for everyone. Some people are in hard-to-publish domains, some people are bad at execution (so they keep missing deadlines and timely results), some people suffer from the endless workloads and low pay. I have observed successful and unsuccessful Ph.D.s, happy and sad Ph.D.s, funding-rich and funding-poor Ph.D.s, and healthy and unhealthy Ph.D.s; from my observation, a Ph.D. needs to be good at communications, good at grant writing, good at conducting research, good at public conversations to translate your hard-core research into an easily understandable version, good at recovering from failures and rejections, and good at obeying people but still using extra time to do innovative work. As a Ph.D., people always love to ask if you have extra time, do you want to do this and that. Be okay with it.

Is it necessary to do a Ph.D. in AI now? -> I still think it’s valuable. It is easier to do AI research as a Ph.D., as research takes time to learn. It’s not an easy task. If I just graduated from college and did several years in the industry, I don’t know whether I would have time to learn and adapt to the challenges, as working while studying is super stressful and hard. Having a Ph.D. also gives you more career choices: for instance, for me, I can do academic work becoming a professor or lecturer (not to mention I applied but got 0 offers), or I can also do machine learning research in a frontier lab in the industry, or be a data analyst or quant researcher in quant, or even start my own company. All these options become more feasible and reliable if you have a Ph.D. degree.

Rejections are hard to navigate, especially in the beginning. I usually feel dealing with rejections needs a kind of mindset: no matter if it’s success or failure, in the end, it’s hallucination. You are still who you are. Be persistent and eventually you will get to where you want.

C. P. CAVAFY wrote a poem called “Ithaka” and there is one favorite line (this is one of the reasons why I decided to go to Cornell):

Thanks for reading.

YH.