Amazing Achievements: CDS Affiliated Faculty Celebrates the Publication of Masters Students’ Research

NYU students are known for being smart, resourceful and, most of all, for amazing us with their achievements. Just one of which is the recent publication of several student papers by students Cole Smith, Andrii Dobroshynskyi, Diwen Xue, Willie Ye, Yueping Wang, Ryan Moriarty, Howard Ly, Ellie Lan, Yijun Tian, Waii Ng, Jialiang Cao. We spoke to CDS Affiliated faculty Suzanne McIntosh, who acted as a co-author and teacher to these students, about the process.

The following has been lightly edited for clarity.

CL: Getting the opportunity to have your research published as an MS student is pretty unique. How do these opportunities come about for the students?

SM: The MS degree program does not require MS students to publish, however, I encourage my students to complete course projects of significance and worthy of publication in peer-reviewed conferences and journals. Whether they publish a paper or a poster, listing a peer-reviewed publication on their resumes and LinkedIn profiles is a differentiator in the job market. Publishing demonstrates and affirms academic maturity, writing ability, successful collaboration, and an awareness of the state-of-the-art.

In both graduate level courses, students implement and demonstrate self-designed Big Data analytics projects. They also write a paper describing their project using ACM or IEEE conference paper style. Once that first version of the paper is written, it doesn’t seem like such a distant goal to complete a paper that can be submitted to an ACM or IEEE conference.

Early on in the course, I introduce students to the conferences most relevant to the types of projects we’ll be developing. I think this motivates students to aim for a project worthy of publication. Through the years, many students have taken the opportunity to work with me to identify conferences that would be the best fit for their work, and then we work together to write the best paper possible.

CL: Which courses did the students come out of? Are there particular courses in the MS track that result in this type of work?

SM: Students who recently succeeded in having their papers accepted for publication completed course projects for either the Realtime and Big Data Analytics course, or the Big Data Applications Development course. Although both are graduate level courses, I have had several undergraduate students successfully complete the courses. In fact, three out of four student research teams who published papers in the past year included undergraduate students. With so much interest from undergrads in Big Data analytics, the Courant Computer Science department started offering Processing Big Data for Analytics Applications — an undergraduate version of the aforementioned graduate level Big Data courses.

CL: How do you guide your students through this process and what advice do you give them?

SM: My first piece of advice is to read, read, read. Once students select the target conference, I encourage them to read as many papers as possible from the past couple of editions of that conference. In particular, they should keep an eye out for papers that overlap their own work in some way. Read those papers and answer the question: Is our project making a contribution beyond the papers previously published at this conference? If the answer is no, then the likelihood of the paper (as is) being accepted is low — the team needs to decide whether they will enhance the paper by conducting more experiments that generate contributions beyond the state-of-the-art.

I encourage students to think from both the author perspective and the reviewer perspective. From the author perspective, we might ask ourselves: How do I target the paper to the chosen conference? Is there additional material that would strengthen the paper? From the reviewer perspective, we might ask: Is the paper well-organized and easy to follow? Is the writing clear? Have I spell-checked and grammar-checked (a common pitfall)?

Experience gained as a reviewer informs the advice I dispense to students. Reviewers have the final say — even if there is a rebuttal phase, and even if the reviewers assign a paper shepherd, reviewers still have the final say. As we are writing the paper, I remind students that the reviewer is their customer and we need good organization of the material so we don’t confuse the reviewer, i.e., we need a clear path through the work. By doing this, we assure that our ultimate customer, i.e., the future reader of our paper, will enjoy reading the paper.

I advise students to place a crisp list of contributions up at the front of the paper so the reviewer knows what’s coming. In so many papers, this key step is missing. As authors, we sometimes struggle to identify and crisply state what is novel in the work — but that struggle is necessary to lay the groundwork for the rest of the paper.

Aside from the actual writing, I ask the students what their goal is. For example: Is it to publish something as soon as possible in any conference? Is it to publish in a top-tier conference? If it’s the former, then we look for the nearest due dates and get to work refining the paper. If it’s the latter, it will likely take more time so we can develop additional experiments to strengthen the existing work, study the results, and write them up.

Cole Smith, a co-author on “Quantifying Local Energy Demand through Pollution Analysis” also commented on the achievement of having his work published by saying,

“We are enthralled to have presented our research at LOD 2020 this year. The issue of climate change is critically important to me, so it’s a great honor to have jointly made a contribution to the field of Big Data Science, and of Climate Research. I’m excited to see similar works published in the future.”

A complete list of papers and co-authors can be found below:

Quantifying Local Energy Demand through Pollution Analysis.

Cole Smith, Andrii Dobroshynskyi, and Suzanne McIntosh.

International Workshop on Machine Learning, Optimization and Data Science (LOD 2020).

Tuscany, Italy.

July 2020.

Implementing a Domain-Independent Framework to Detect Suspicious Review Patterns.

Diwen Xue, Willie Ye, Yueping Wang, and Suzanne McIntosh.

IEEE BigData 2019.

Los Angeles, CA, USA.

December 2019.

Deal or No Deal: Predicting Mergers and Acquisitions at Scale.

Ryan Moriarty, Howard Ly, Ellie Lan, and Suzanne McIntosh.

IEEE BigData 2019 — International Workshop on Big Data for Financial News and Data.

Los Angeles, CA, USA.

December 2019.

Geek Talents: Who are the Top Experts on GitHub and Stack Overflow?

Yijun Tian, Waii Ng, Jialiang Cao, and Suzanne McIntosh.

International Conference on AI and Security (ICAIS’19).

Brooklyn, New York, USA.

July 2019.