On the evening of December 8, 2016, exactly on 5:01 PM in Eastern Standard Time, the photons of Cornell Engineering acceptance letter struck my eyes. For me, it was a confirmation, a confirmation for an unexpected success of a rebellion.
I can imagine what would happen if things were the other way around. There would have been some discussions in teachers' office. "Another failed example, just like John Shi two years ago," commented Mr. Alpha. John was the leader of the student union, founder, and president of the robotic club and a film-making club. He got the perfect score in IB, 45 out of 45. He graduated from my high school two years ago, with offers only from HMC and UC Berkeley. Although HMC and UCB are two schools out of reach for many students, it should be the safety school for him. His failure foreshadowed this year's college admission in our school, especially Class of 2017 generally less fond of humanities than sciences.
When I heard John's story, I was only in tenth grade, enrolled in the high school for just one semester. Rumors began to spread. Our school was generally perceived by outsiders as a school of excellent extracurricular activities. It was indeed true. Besides tens of student organizations in our school, the school also help organized a trip to Dabie Mountain for teaching poor children and tree planting activities in Inner Mongolia. Also, some of the great students can even organize some activities that attract media attention, for example, the annual charity. However, it was just one side of the story. Its implication had put great disadvantages on students like me. The logic is simple: they are only good at extracurricular activities; they are poor in math and sciences. You can argue, or even protest, but before protests become effective, it is a fact, no matter if it is an acceptable one. Now, it seems like the already bleak future of my college admission future was added another opaque layer of uncertainty.
About the same time, it was also a time of decision and determination. After a semester of exploration in the school, I became a member of CCC, a programming club, and a member of the student union in the publicity department, or MiniTrue in NewSpeak. I also participated in Stanford Math Tournament with 7 other students in Grade 11. Three of the group members, Henry, Ben, and Lily, strongly recommended me to join the club of Computerization which devoted to developing an online platform to serve all the students. Due to increasing time constraints thanks to the evil College Board, I can only afford to choose one of them. I already became bored with and decided to leave CCC, which only creates computer animations. Therefore, I had to choose between Computerization, a club of nowhere but I liked, and student union, a "student" organization of everywhere but I had an unpleasant history with.
I disliked the student union for a few reasons, one of which is the difference between expectation and reality. The expectation from the Gettysburg Address told me that the union should be of the students, by the students, for the students, but the harsh reality speaks for itself that the union is of the school, by the establishment, for the elites. I did not experience the early history of it when innovation is still abundant or possible. I was admitted in the school when the student union was in its fourth year when many innovations in the earlier time were repeatedly proven successful so they became mandatory. As a result, the delegates in the union were too preoccupied with these established sequences of events, ignoring possible new innovation. When I was admitted the member of the student union, I wrote a lengthy letter to the head of MiniTrue, describing a possible platform for the publicity as well as for the discussion of all the students. I recommended the use of machine learning algorithm based on students' discussion that could not only target students of different interest more effectively, but also guide future activities planning. The plan was acclaimed and was only acclaimed. I never heard of any mention of it until I submitted my resignation letter. If ignorance of technology was slightly better than bearable, then the betrayal was intolerable. One month before the annual charity event, I was asked to update the charity website that served to sell the tickets. I thought that it was a great opportunity to demonstrate my programming skills and to have a say in the department. When I devoted hours to it and upgraded it with flat and responsive design, I received the instruction on the class monitor meeting: please recommend students to buy tickets via class monitor instead of the website so that the paying service provider can share less revenue. Once again, technology was ignored.
However, I was not an idealist. If I were, I would quit when they first ignored my plan. The power and resources of the student union are highly desirable when you want to organize some activities and events. At that moment, Computerization was still in its infancy. Nobody in my class has ever heard of it, nor does it have any real influence in the school. It is doubtful whether it can have even 10% of the momentum compared to that of the student union. Idealism defeated pragmatism when I decided to bring a new club into prominence. Pragmatism defeated idealism when I gave up the title of student union member to have the real power to make the change. When I first joined, they were trying to promote a modified version of Moodle to serve the homework assigning website. I persuaded them to replace it with a self-developed platform and spent about two months to create it, which we later named SAM, System of Assignment Management. With another year of work, SAM expanded to have 40% of student population and Computerization trained grouped the best software engineers on the campus into a great team. Now that I am looking back into the history, I found that the casual moment when we finally decided the name of the platform is SAM is actually the ritualistic moment when my idiosyncrasy had been merged into the club. By the end of June, when the old leaders Henry and Ben were ready to concede the leadership to me, SAM was already a Sam project in which I contributed 99.9% of the code. The most intensive work of the club has been transformed from promoting an existing project to constantly making great things to surprise the world. That shift in the workflow is indicative of the shift in attitude. Sam was no longer a businessman, he became an engineer; Computerization was no longer an activist group; it became a research and development organization. We no longer cared about the actual statistics of active users; we cared about the quality of our service and the difficulty of the research project.
SAM on GitHub
Concurrent with the fundamental change in Computerization, there was an ongoing battle: mathematical modeling. It started when I signed up for HiMCM 2015 in 2014 fall and started to teach myself calculus, but intensified when Calin, Tim and Samuel and I took a formal preparation course starting from basic mathematics and tutorials of MATLAB ending with some smart algorithms that we ended up using. Now with the advantage of hindsight, I can argue that the experience in Computerization and mathematical modeling shaped my nerdy life, and gave me the sense of Nerd Pride.
All the Math Books I Have
Before these activities, I was highly unsure about my future. Instead of a concrete plan, for example, for the training of a future software engineer, I only had some vague ideas: probably some position in Google, probably in a startup company, etc. I even did not exclude the chance of being the founder of a startup myself, just like mostly what I did in Computerization. The harsh reality in user statistics pushed me to the other side. While we were promoting our platform, teachers liked it and forgot it; therefore, the user coverage was always around 40%. In the 11th grade when I was fully in charge of the club, I always wanted to throw me into earlier times when all I needed to do was to solve complex technical problems one after another. I became increasingly aware of that I am an engineer, not a salesman or businessman. I am the one who builds great products but not necessarily has excellent skills to promote them. The success and failure of HiMCM even strengthened by belief. I was able to incorporate all different mathematical and computer science knowledge, including multivariable calculus, statistics, linear algebra and genetic algorithm to analyze and optimize the traffic on highway, but eventually failed in terms of presentation when later proofread of the thesis suggests that our writing may lead others to believe we only tangentially solve the road rage problem required in the question set. We got Honorable Mentioned, which is only better than Successful Participant, a prize that everyone excluding cheaters can obtain.
It is confirmed. Sam is a nerd, not the type of businessman favored by the school.
With the nerdy spirit awakening, I found increasing uneasiness with my old ally-turned-opponent, the student union. The student union held its indirect election in May, via electoral college from almost randomly chosen electors to decide committee of the next generation. It would not result in the disagreement between electoral college vote and popular vote because the latter concept does not exist. The nomination process was controlled by the school itself, which involves some opaque backroom dealings that sometimes surprise all the participants. As a result, there is no third-party (or even second-party) nominees or independent candidates. The final decision is also hilarious: electors choose 10 from 12, which means that a school-approved candidate has 83.3% chance of winning. I witnessed both the student union election and the US presidential election. I also witnessed the decline of their quality. Excellent promotional videos were replaced by perfunctorily made propaganda; professional lectures from academia were reduced to dogmatic teaching of success theory; more importantly, work was rebooted into talks, mere talks on incomprehensible prose with unsubstantiated substance on ambitious daydreams. Once they invited a famous guy to give a lecture on his successful life. In order to obtain a satisfactory statistics, they forced all the student leaders to attend and disguised it as club president training to trick more. Half a year later when they were summarizing their jobs, they unashamedly claimed that they were guided by the unchanging glorious ambitions.
It turns out that there is a high correlation between being a minister in the student union and being a manager in the student film. There is even a stronger correlation between being a minister in the student union and majoring in economics or other social sciences. If my personal failure in Computerization and mathematical modeling had already blocked that path for me, the actions of the student union had destroyed the reputations of those areas. I was only left with one choice, engineering. I chose it and I chose it with pride and determination. I felt that I had a built-in nerdy spirit, but previously, it was private and silent. However, it was the time for a change.
After the first term of the 11th grade, I made the first choice: the type of summer program. Many of the students went to the summer program to have a university, leadership, or just a party experience. The content they have learned, on the other hand, may not be very important. It is mostly evident from the sharing of their experience after the summer holidays. While I talked a lot about the math I had learned, they were talking about how hard they had worked and thus a wonderful university experience well bounded by expectation. I was not fully aware of the negative side of generic university summer programs, but at the time of the application, I simply found math programs were far more interesting than those which admission depended on some random essays, useless recommendation letters or even nothing. The math programs that I applied all required a written test of competence. While essays are required, they were mostly used as tie breaker according to my speculation, because I did so bad in essays but excellent in math tests. Not only did the mathematical questions presented in the tests interesting and worth a bad sleep for consideration, but also the test itself was a reflection of my own philosophy: shut up and show me your work. Finally, I went to MathILy to spend five weeks deep into abstract mathematics that most businessmen never hear of: fractal, topology, complex analysis, etc.
The Most Interesting Question of MathILy Admission Exam
On March 11, the triumph from the old era eventually arrived: I finally obtained 2310 in the old SAT. Suddenly, I became free of any major standardized test and was unexpectedly granted tons of time for independent exploration. Others discouraged me to even to touch the boundary of online activity analysis due to its tremendous difficulty. True, the time cannot be wasted or I would lose an edge in college admission process. What an excellent and snobbish wisdom! I once again rebelled against the convention and chose to follow my heart. It all comes from an incident three years ago in middle school.
In the middle school, the gossip was flagrant. There is one guy, who I preferred to call Mr. Beta for the sake of privacy, who was very good at psychological analysis. Sam was exactly the opposite, but with an additional knowledge of computer science. He claimed that he could know whatever I thought. I was unwilling to admit my apparent failure, so I responded: my computer can. The childish verbal exchange of fire would sink into the unknown corners of history very soon if happened elsewhere, but not for me. The claim was not entirely unsubstantiated: Google Brian learned what is a cat without any human interference; machine learning was developing fast and violently; artificial intelligence was on its rise. The only, but tremendous, difficulty for me was to apply the current development in the fields to the analysis of online activity to achieve a similar result of facial expression analysis.
The incident occurred in the summer of eighth grade. Then it was a year of deadly preparation for High School Entrance Exam that nearly killed me. In tenth grade at the high school when I had not yet overwhelmed by SAT, I taught myself calculus to build a mathematical foundation for that work. A year later when I finally conquered the SAT, and when I was rich in mathematics after the training for mathematical modeling, I was ready to re-take the challenge. Neural networks, activation functions, regularization, convolutional neural networks, recurrent neural networks, long-short-term memory, etc. Tons of new words flowed through my mind. It was like figuring out the core of calculus without the knowledge of the sine function: imagination, guess work, and inevitable limitation. But in June, the clouds were clear when my neural network predicts the future evolution of interests with an 88% accuracy. The job is done.
Code for Constructing Convolutional Neural Networks in Research
I had spent most of my life being a lonely nerd, but at MathILy, nerds rule. All the silent nerds at school became outright. We talked about mathematics instead of stupid fashion or latest gossips. We were sad for bugs in our proofs instead for a decline in popularity. We made fun of others' name by relating them to mathematical definitions. We all laughed when someone was secretly using second order ordinary differential equation just to find the generating function for Fibonacci sequence. It was the greatest time. All the social norms were gone and replaced by the genuine interest in pure and applied mathematics.
When my high school started two years ago, I was in a period of uncertainty. But when I was beginning to figure out I am indeed a nerd, I started to strive for being great: learning calculus before anyone else in the school, creating software better than anyone else's in the school, doing computer science research powerful than anyone else in the school had ever researched. The price could not be taken lightly: overly reliance on coffee, some white hairs, and a great disadvantage for college admission: a nerdy spirit that is virtually impossible to conceal. Yes, it is the time for college applications.
On July 30, a five-week mathematics course at MathILy had ended; two days later, an eight-day college campus visiting trip started. The list of colleges to visit had been finalized in May, when I was still undecided but narrowed my ED choices among Cornell, Duke, and Penn. Initially, the NETS program at Penn drew my attention since it provided a highly integrated course for big data analysis. Duke then replaced Penn when I found that only Duke does not discriminate against male applicants among three schools. However, I eventually chose Cornell when I detailedly compared their courses in computer science as well their education in mathematics. Therefore, I almost already made up my mind to apply for Cornell in ED. However, for the sake of increasing success rate, I initially decided to apply for College of Arts and Sciences or even College of Agriculture and Life Sciences which has an information science major.
I first researched extensively to distinguish between computer science and information science. I hoped the research could prove that CS and IS are equivalent or IS is better so that I could have an easy option. The result did the opposite. Now I was left with only one option: College of Arts and Sciences, where none from my high school had ever been admitted into. While I finalized my own research, my parent came up with a completely different idea: applying for NETS program in Penn Engineering. We all knew that Penn is significantly harder to get into. We also knew that I was willing to take the risk since I would be very comfortable to be admitted into UC Berkeley in case I failed the ED. Thus, it was not the choice between risk and safety. My mom's message came in during the last week of MathILy course. She cannot elaborate on her points simply by chat, so we all agreed to have it settled after those campus visit. Suddenly, the visiting, which was only supposed to be a confirmation when I initially chose Cornell, became a contest and a source of debate. The topic of the debate was fairly simple: Cornell is a better choice than Penn. The affirmative side stands the nerdy Sam; the negative side stands, as I later discovered, the whole world. The evidence to either prove or disprove the claim will be carefully collected on August 1 and August 3 when we would visit Cornell and Penn respectively as scheduled.
On August 1, we first arrived at the information session of College of Arts and Sciences. It was the first info session, so I was busy taking down notes for the source of my application essays. After the information session, I specifically asked the lecturer to tell the difference between computer science and information science to ensure my judgment is correct. It is. He said that information science is more of an integrated subject, while computer science is more rigorous and theoretical. It is basically another version of the statement I found on Reddit: information science is for those who fail computer science. While I knew I am more of a nerd less of a businessman, I have to study computer science instead of the information science to sharpen my ability to make great things. Now the easiest path had been completely eliminated. I have to take the risk. Then we went to the info session of College of Agriculture and Life Sciences, in which I almost fell asleep. Besides these two colleges, I also registered for the info session of College of Engineering, although I thought I had almost no chance of getting in.
The registration of this info session, initially for utilizing the time between an info session and campus tour, proved to be one of the most important decisions from the hindsight. When the lecturer walked into the auditorium, his first sentence was not "Welcome to Cornell" or "Welcome to Cornell Engineering". Instead, it was "Welcome to organic chemistry." From that instant, I knew it was the place I would like: a bit nerdy, openly nerdy, and proudly nerdy. In the middle of the session, when the lecturer was explaining who should apply for Cornell Engineering, he stated the motto of engineering students: "I love math and science. I love problem-solving." It was a direct hit since even the nerdier CMU did not say these words in such an open manner. OK. College of Engineering was significantly harder to get in. Boys were discriminated the most in the application. However, I would love to take the risk. I remembered that I took two potentially life-changing gambles previously. The first time was in December 2015, when I refused to take the new SAT and concentrated all my effort on the old one. The second was in May 2016, when I did not renounce when weeks of research in machine learning only produced meaningless data. I narrowly won the previous two. Now there should be a third one.
However, it was just the interpretation from my perspective. My mom was still thinking of ways to persuade me to apply for Penn. From her perspective, Penn had a better reputation among the Ivy League and NETS had a better course system that gave me a well-rounded education. She also believed that the business environment at Penn can convert a nerd like me into someone else. What she appreciated was exactly the reason I preferred Cornell to Penn, so our discussion was not even on the same page. On August 3, the disparity became more obvious. I had to admit that Penn did a very good job in the information session: well-controlled light effect, wonderful speech with historical and cultural significance, and discussions on the imagination and beautiful dreams. However, they were exactly the same things I did not care about. To make it worse, it reminded me of the unsubstantial and platitudinous speech delivered by the ministers of the student union. I did not want to be a member of student union; I belong to the nerdy Computerization. The same logic should apply to Penn and Cornell.
For me, the decision was already made, but it took another three weeks to debate on which school is better. Everyone tried to persuade me: some believed that I deserve a better school as if Cornell Engineer was easy to get in without two successive AMC 1%; some repeated the same talk on reputation; others focused on the courses. I was not persuaded, just like I was not persuaded to remain in the student union, to research on simpler topics instead of complex ones, or to be a businessman and not to be a nerd. On August 24, I produced the first salvageable draft of Cornell Engineering Supplement essay. In the essay, I proudly wrote about the middle school incident and the horrible research experience without help. I mentioned the loneliness as a nerd in an overwhelming environment for future businessmen and businesswomen. I quoted exactly the same words "I love math and science. I love problem-solving." The act of writing the essay itself and the content of the essay perfectly demonstrated my determination, and more importantly, my Nerd Pride. There were no more debates. The decision was officially finalized.
Three months later, at 0:01 AM in Beijing Time on December 9, six hours prior to the official release of Cornell's decision, I was still unable to sleep. I woke up from a bad sleep at 4 o'clock, then started to read mathematical proofs to relieve my stress and to pass the most nervous two hours. At 6:01 AM in Beijing Time or 5:01 PM in Eastern Standard Time, I finished reading the proof of Gaussian Integral and refreshed the web page.
Everything has a beginning has an end. I prefer to end with the ending of my Cornell Engineering essay: "My journey in artificial intelligence has only just begun and I hope my next steps in that journey are taken at Cornell."
Letter of Acceptance from Cornell Engineering