GRADUATE SCHOOL
[CMU MISM-BIDA] Attending Graduate School at CMU in Fall 2021
At Hamburg Hall, greeting tartans on CMU campus!
WINTER IS COMING to Pgh! BUT Spring came earlier…
Preface
It’s spring break now! So I had some time to be back on Medium… Sorry for all that time (calculating seconds or MONTHS)! My first REAL in-person semester at CMU. This means going to all school classes and meeting everyone but with masks. To be frank, I am pretty satisfied with CMU’s policy for the pandemic. In addition to keeping everyone safe, we still get to a taste of what school is like. Many of my lecturers this semester really shines in an in-person setting. I will give more details on that later in the article. My last article about my adventure can be found in the hyperlink: [CMU MISM-BIDA] Attending Graduate School at CMU in Spring 2021 (use wisely)
Carnegie Mellon University
Heinz College of Information Systems and Public Policy
Master of Information Systems Management — Business Intelligence and Data Analytics (MISM-BIDA)
Overview
So I am currently drowned in job search where I got a little sick mentally and then it affected my physical health. But I think I got better now after spending some good night sleep and doing yoga and jogging? I hope that no matter why you decided to read this article: STAY SAFE AND STAY MENTALLY HEALTHY :) (This smile is for you) Again, this article will be simply about “graduate life,” and will be posted on my publication: As a Graduate Student in Data. If you are interested in reading more about my graduate experience, feel free to check it out! I will go over my experience in the courses and life at Pittsburgh! You will, however, not see any internship-related parts in this article. (future articles to come!) Let’s get started!!
Fall 2021 Class Schedule
For those who are in the BIDA program, AEI (94–834) and ABA (95–866) are BIDA-specific cores that should be taken in fall (Cases may vary per school year). Other Heinz cores are JAVA (95–712), DT (95–722), and DMUU (95–760). I took Java in fall is because I did not take it in spring 2021. Advice: For future MISM students, please consider taking it in your 1st semester… This semester is my second semester. I decided to take the most heavy-loaded courses: IDL!!! And also, the FAMOUS Terry Lee Data Structure.
The following are the courses that are on my schedule:
- 11–785 Introduction to Deep Learning: 12 credits (Full Semester)
- 17–683 Data Structures for Application Programmers: 6 credits (Mini 2)
- 94–834 Applied Econometrics I: 6 credits (Mini 2)
- 95–712 OOP in JAVA: 12 credits (Full Semester)
- 95–866 Advanced Business Analytics: 6 credits (Mini 1)
- 95–722 Digital Transformation: 6 credits (Mini 2)
- 95–760 Decision Making Under Uncertainty: 6 credits (Mini 1)
11–785 Introduction to Deep Learning A+
Attention is all you need. — I just love the paper title!
The famous IDL course was taught by professor Bhiksha Raj from the LTI. 11–785 is the PhD version and 11–685 is the Master’s version. You might be wondering about the difference: 11–785’s last assignment will be a research project while 11–685 will be a homework assignment. I chose to do the research assignment with online version (I had a conflict in mini 1), so 11–785 section B was registered!
Some pros of 11–785 is that it is so well structured and it goes from easy to hard! (For me it’s actually hard hard lol) Starting from the most basics MLPs to the more advanced topics in DL. It is really “an introduction” and I would say it’s a must-take if you are aiming for a data career. Who doesn’t use DL now? Well actually, many companies use DL, but not all of them. (my apologies…) Here are some of the topics that is covered (4 pillars):
- MLP/Backpropagation
- CNN
- RNN/LSTM/CTC
- Transformer/GNN/VAE/GAN
If you are looking for the cons, it is definitely the time invested in it and also the effort of trying to figure out how to implement homeworks, finish the weekly quizzes, etc. (Please refer to the working hours above) There are also lots of papers to read. To be more specific, at least 1 paper per week for the quiz and at least 1 paper per homework. PAPERS!!! For me, it took me more hours than the 22 shown above but at least I got a decent grade in the end!
The format of fall 2021’s 11–785 was in-person and literally I could sense Bhiksha’s humor throughout his lectures. Many people think that Bhiksha is a strict guy but he is actually quite nice and friendly. I remember on Halloween, he asked all the TAs and himself dress up in costumes (refer to the following video if you would).
Another component that shines in for the course is Kaggle competitions for all homeworks! Note that AWS credits will be passed out but I seriously recommend Google Colab Pro! For these competitions, it is when you try and find that “things don’t work”, then you will come back and face your own demons. It is really tough when you are trying to cross an A cutoff and you were just 0.005 below it. This part took most of my time tuning, searching for ways to be better, and simply training my model. I remember going to a TA office hour asking why my model trained 2 minutes per epoch while some of my classmates were running 10 minutes per epoch? I got a reply that “well if it worked then go for it!” This was the one that worked in the bottom picture. (Yeah! Top 10!)
If you want to learn more about this amazing course, please see the link below. Also considering following my Instagram for more up to date posts!
17–683 Data Structures for Application Programmers
You DESERVE an Apple! — Terry Lee
From one in a million, it’s oh Terry! Professor Terry Lee is out of my mind! Some guarantees for this course:
- A start off to know more in-depth about major data structures
- Be very good at motivating yourself with Terry stuff (more on that later)
- Attending a Vegas show at CMU
I have taken data structure before back in SCU but I still did not really know how to use each one at the right time. Terry was like a guide to all that illustrating with examples in life. If you had the chance to take it, you will definitely love him. More on the topics covered:
- Big O
- Array
- ArrayList
- LinkList
- Stack
- Queue
- Bubble Sort, Selection Sort, and Insertion Sort
- Recursion
- HashTable
- MergeSort and QuickSort
- Binary Tree
- TreeMap and TreeSet
- Heap and HeapSort
About the Terry stuff, it is like every data structure has an iconic figure in life. Every time I see something related: “OH TERRY!!!” pops out of nowhere (seriously). One example is on stacks as of Pringles, the concept of LIFO as of popping the top and get what was put in last. OMG!
Terry also loves to give out apples when somebody answers a question in class. So if you are taking his class, get an apple! I got my by participating a LinkList demonstration which was really fun. (see pic on my Ins: https://www.instagram.com/p/CVmETiZtftu/)
For the loading of each homework (in Java), I would say “the recursion one” was the most devastating of all (for me). After looking back on the assignments, I thought it was manageable and also well designed. Each one focuses on building one data structure and will be a pain if you don’t have the logic or coding sense. Some of my friends really struggled. I would recommend this course to whomever wants to know how to better code effectively with the right data structures.
94–834 Applied Econometrics I
Take AEI if you are looking to learn more in metrics and do experiments. The textbook used is “Mastering ‘Metrics — The Path from Cause to Effect.” Edson Severnini, the professor, LOVES the subject he teaches so much. He denotes himself as an economist which I thought was really cool. In 2021, his teacher in UC Berkeley Professor, David Card, won the Nobel Prize on labour economics, which was fascinating! (see link)
We used STATA for each homework. Though I am not a fan of the software, I guess getting to know another tool is not bad. Since this is the first of a two-mini series course, the topics covered mostly overlaps some basic concepts in statistics. I did not take Applied Econometrics II but if you are aiming for roles like Data Scientist, Analytics (see my article on different roles in data) — This may be a good course choice.
- Randomized Trial
- Regression Analysis
- Instrumental Variable
95–712 OOP in JAVA
System.out.println(“Hello World!”);
As you guys may have noticed, Java is huge at CMU. Let’s find an explanation on Google: …
We found this in the 8th Little Brags of Big Ideas:
In 1994, alumnus James Gosling became the “Father of Java” when he invented the Java computer programming language. Java has now become one of the most popular programming languages in use today.
Now we had an answer! I always get questions like why do we have to study Java if we are doing data? Well, my answer for this is that lots of OOP and data structure concepts are wrapped in the Java course and you can be a better coder if you know how to use them. My other answer for this question is that you can try to ask James who studied at CMU and invented the language. (lol)
Now, back to the course, I took professor Marty Barrett’s version, where most of the Java courses are taught by professor Neelam Dwivedi. Java consists of 24 labs, 14 quizzes, 3 homeworks, 3 midterms, and 1 final. Marty said in the first class that you are expected to know how to code effectively in Java after this course but they are all easy for me. While some of my friends struggle in the course, I love the course and thought it was pretty straight forward.
Students worked on several assignment projects which different applications are built. Most of them are build on one another. The thing about CMU coding lectures is that you don’t get a standard answer but only suggestions on what you did wrong in your previous homework submission. That’s part of the challenge! Java is also a prerequisite for Distributed Systems so I would definitely recommend students to take it in the 1st semester (saying it again).
95–866 Advanced Business Analytics
Microsoft Excel Solver: GRG Nonlinear
ABA can actually be renamed as Advanced Statistical Theory: Maximum Likelihood Estimation (MLE). All the topics are wrapped under the umbrella of MLE just with different data distributions. The course also requires lots of operations in Microsoft Excel, so if you are an Excel fan, you will definitely be happy (I’m not)! I really enjoyed one of the guest lectures that was on recommendation systems (I love RecSys). The lecturer was from Google Brain and she also gave us great insights into her research career in data!
- Survival Analysis
- Count Data Models
- Choice Model
After talking about the pros, I would go over the cons. To begin with, the homeworks might be very challenging since methodologies covered in lecture might not be useful in the homework. Furthermore, it was a pity that the course was later moved to Zoom. I was expecting an in-person course when I received an email form the professor.
Since this is a BIDA core, I think what Heinz was thinking is that Excel is still widely used in field and once you know the concepts it does not really matter what tool you use. But it makes me wonder if there are other more important topics than MLE that could build on Statistics for IT Managers (see my review of the stats course in this article).
95–722 Digital Transformation
Michael D. Smith is another professor I love so much! He got his degree at MIT and he is always so passionate to share about his findings in life and while he is teaching. One word to describe his class — Chill! And probably lots of discussions! People are always raising hands and commenting on things.
For each assignment, students work in groups to write on different topics in digital transformation. Topics that were covered include: Structural Change in the Music Industry, Digital Transformation in Security, Digital Transformation and Higher Education, etc. Though this is more of a business-oriented class, I felt like we had the opportunity to read more about what is happening in the world. Since this is a core, I would definitely recommend taking Michael’s version of DT.
95–760 Decision Making Under Uncertainty
Microsoft Excel Solver: Simplex LP
DMUU can be renamed as Optimization, which is Shorter and easier and more understandable. So if you notice that both ABA and DMUU uses Microsoft Excel Solver, there must be a difference, right? Well, yes! The main difference between ABA and DMUU is that ABA uses “GRG Nonlinear” while DMUU uses “Simplex LP”.
Again, this is a Microsoft Excel class. My professor that semester was David Choi, an angel who always wants to give his students the best learning experience. He constantly ensures that his powerpoint are on point and that the Zoom audience can see his writings (Yes, he writes out the formulas in class). The recommended textbook is Spreadsheet Modeling and Decision Analysis: A Practical Introduction to Business Analytics. We learned about lots of linear programs:
- Linear Programs (LP)
- Network Flow LPs
- Two-stage/Stochastic LPs
- Integer Linear Programs (ILP)
- Simulation
I meet an alum who said that he thinks this class is super useful and that he utilizes David’s optimization methods in retail. I would conclude that this is another extended branch of statistical methods that Heinz wanted the us to be equipped with.
Pittsburgh Life
Some of my audience only wanted to see this part of the article, so I will leave you to all the pictures! If you want more on Pittsburgh life, feel free to check out and follow my Instagram.
Newell-Simon Hall
Hamburg Hall
Gates & Hillman Centers
Millie’s Homemade Ice Cream
Heinz Field
ASCEND Pittsburgh
Carnegie Museum of Natural History
Tepper Quad
Moore Park
Jared L. Cohon University Center
College of Fine Arts
Epilogue
1 year back in the States! I would conclude this year as finding a new way to know MESELF. Though there are a lot of geniuses at CMU, I also met a lot of stupid people along the way. I found that it’s ok and it’s fine to think that they are idiots or selfish nerds (wut?). They are just people. There are people that insisted on being wrong and you can not change that. It is just losing the game in another way. Why do people like to overestimate themselves? Why do evil people get to live longer? Life is a rollercoaster ride and after all that, I said to myself:
Why not enjoy the journey?
My 2nd official semester at CMU has been DEEP LEARNING! My next semester will be my last semester, which is Spring 2022. Thanks for joining on the ride as I will dive into more of my school life at CMU next time. If you have learned something or have a similar experience, feel free to comment below! If you have more to discuss, you can contact me via LinkedIn, Instagram, or Twitter. Also, feel free to follow me on Medium for more graduate life articles to come!
Previous Blogs (Let me keep it as 5s)
- [CMU MISM-BIDA] Attending Graduate School at CMU in Spring 2021
- [CMU MISM-BIDA] Attending Graduate School at CMU in Fall 2020
- 6 Questions Answered at CMU’s Esports Industry Night
- Pandas!!! What I’ve Learned after my 1st On-site Technical Interview
- Hey Newcomers! Let’s Peek into Data Science Jobs
