Trials and Tribulations of a Senior, Part I
Long entry because I haven’t written in a while. Most of it is me rambling.
This week I had my Senior portrait taken. I thought it was so cheesy that my mom was making me do this…but it is the least I could do. As soon as the photographer put all the paraphernalia on me (cap, gown, cord, sash), I felt different. I realized that a chapter of my life is coming to a close. Sometimes when I look back I feel like I have accomplished nothing during my College career. But I now realize that it is because at this point in my life, I cannot yet accomplish my dream: to get a doctorate. This Bachelor’s degree means more to me emotionally than it does academically. It represents all of the sleepless nights, all of the hours spent in offices with TAs and professors, all of the struggles, all of the heartaches and tears I put into this career.
My college career has been about improving myself and teaching me a new way of thinking and looking at the world, more than teaching me anything academic that I am ever going to seriously use. I cannot say that I knew how to prove the Cayley-Hamilton Theorem before I came to college, but material like that is water under the bridge.
Adding Statistics as my second major is the best thing I could have done. I got to discover a department that actually cares about its students. It has made my college career so much more enjoyable…and normal. This last year in college with the Stats Department so far has proven to be more rewarding than the last 3 years combined.
On the Grad School application front, things are going well. My grades are good, especially in both of my majors. My GRE scores are really good although I will probably retake them because I think I can do better. I was freaked out taking a Computer Based Test (CBT) (there is an entire branch of psychometrics [my interest] dedicated to this). I have secured three letters of recommendation from professors that think very highly of me. They have been amazingly supportive. That support I could never ever get from the math department and its faculty. The only thing that I have left to do is write my Statement of Purpose which I am hoping to do this weekend. I need to write about why I want to be admitted to Grad School, why Statistics, and why UCLA. I feel so strongly, and so passionate about this opportunity, I can probably write a very strong statement on the first try…and then read it over and instead lament over it for days. My proposed specializations are Psychometrics, Pattern Recognition/Machine Learning, and Data Mining. They all have their pros and cons, and are all very fascinating subjects to me. Psychometrics (measurement) is a branch of statistics that deals with psychological and educational measurement: how tests and scoring metrics are constructed. My interest is educational testing as opposed to psychological testing but the concepts are all the same. Last year I wrote a paper about a metric I conceived that measures educational growth over a grading period called R******’s Combinatorial Growth Metric. It was fun to write, but there are horrible flaws in the theory and I would never attempt to publish it because even I doubt its purpose. Although I was scared to death to take the GRE, I was fascinated at how much more convenient it is, and how much more accurate and concise the testing and scoring is than the paper-based test (again, psychometrics). My second interest, pattern recognition and machine learning is perhaps the most theoretically math intense branch of statistics. Researchers in this field study how to detect patterns in data (data is much more than just a “rectangle.” It can be a picture, a sound, an energy, a map, you name it) and model those patterns to extract their meanings and make meaningful predictions. Signal processing falls into this category. It is possible to, for example, take an audio sample of an auditorium clapping, and pinpoint one person’s applause. Pattern recognition is also used in face and fingerprint recognition. Eventually I will summarize face recognition theory and how we all come from a small basis of faces: eigenfaces. My Optical Mark Recognition project also relies heavily on pattern recognition (image processing) and Bayesian statistics. Machine learning relies on pattern recognition. Some examples of machine learning are robots, and computer based testing (such as the GRE General). Pattern Recognition and Machine Learning are an exotically beautiful mesh of theoretical mathematics (especially combinatorics and linear algebra, my favorites), theoretical statistics, and computer science. Data Mining is the most applied of the three fields. It is basically a combination of statistics and computer science. It basically studies how to obtain data that are not in a “rectangular” form and how to visualize that data so that the layman can understand it. My OMR project can also fall into this category.
I have also turned 22 since I last wrote. I feel old. My birthday was awesome though. I hate cake, so instead my mom made me a bunch of cupcakes in the shape of a cake and put candles in them. It was perfect. For my birthday I got a new phone (LG VX8100) that does all the cool new things: realistic music ringtones, ringback tones, picture/video messaging, cameraphone, speakerphone, and I can even send the pictures I take over the internet to an online photo album. I also got an iPod. I said I would never, ever, ever get one, and well, I finally said “what the hell.” It makes the long 20 minute walks from MS to my place faster, and makes grading less boring.
Working three jobs has taken its toll on me. Grading 200+ papers a week can get very frustrating and infuriating at times. Grading for Statistics has been great though. The classes clearly understand the material and put in effort into their work. Linear Algebra is a totally different story. I don’t know what this professor is teaching the class. But I can clearly tell you what he is not teaching them. It pisses me off that in week 8 there are still students that cannot row reduce. I have never graded a class where so many of the students put so little effort into their work: skipping half of the problems, showing NO work, cheating, not stapling their 10 crumpled pages, not putting a name on the paper, turning in homework late (repeatedly). It is insane. And the fact that the professor allows his students to turn in their homework late (repeatedly) for every damn reason is even more infuriating. Sunset is still a great job. So clearly, the sore spot is the Department of Mathematics.
There is something else that has taken a toll on me but some things are not worth discussing in great detail. What I will say is that I never thought four (actually five; the fifth I don’t even know) men in their twenties could be so low, so shallow, so selfish, so idiotic, that they would let somebody do something so evil to me…and then not tell me about it. To make it worse, these guys were so spineless somebody else had to relay the information to me. The story is that it is a joke gone horribly wrong but I am still suspicious that the whole thing was intentional, and I am pretty content with that feeling, and I am confident in my personal accusation. What seals my suspicion is the fact that I never even got an apology, and their reason for not telling me was beyond pathetic. I am proud of myself for how I handled it when and after it happened, and quite frankly, I’m pretty surprised at how I handled it. I was never even close to caving in and so they should get it through their thick heads that I will never cave in, especially not to them. And this is the reason why they are now pissed off at me. And that makes me proud. I was right about them, and I have no reason to ever question that feeling again. I want to thank my coworkers at Sunset for their support when this happened. As they say, what doesn’t kill us makes us stronger. But now that I have put my feelings in writing, I feel that I can finally move on, and not need to write about this again.
Classes were very rough but now that I have dumped Artificial Intelligence I have had more time to dedicate to Pattern Recognition and Statistical Programming. The problem in Pattern Recognition is that up until now I have been taught frequentist probability which is the old and conservative way of doing things. This class requires understanding Bayesian probability which to me is a totally different language, but fascinating nonetheless. The theory required in the class is amazing. We have concepts from Analysis, Combinatorics, Linear Algebra and even Topology. I had a very difficult time understanding methods of dimension reduction such as Principal Components Analysis at first because it is very difficult to even attempt to visualize a d dimensional vector space containing data. This is where the eigenfaces come in, and my project that can best be described as a disaster. Then we moved into non-parametric methods of probability distribution learning such as using Parzen windows and the nearest neighbor technique to classify data into 2 or more mutually exclusive groups.
That’s the latest. I am hoping to post my Statement of Purpose when I have it done because it should be interesting. Thanks to Wikipedia for making my blog interactive 
.
