911±¬ÁÏÍø

Soham De is a Ph.D student at the iSchool.

When you meet someone who doesn’t know about your research, how do you describe it?

Broadly speaking, I study social media algorithms. My research at present focuses on two
specific types of algorithms: content recommendation and moderation algorithms. The first type refers to algorithms that determine the content on your feed while the second consists of algorithms that intend to deal with the censoring of toxic or misleading content. 

Who is the faculty member working closest with you? What are you learning from them?

I work closely with my faculty advisor, Martin Saveski, who is such a great and supportive
mentor. Martin has taught me everything I know about research. To name some specifics, I’ve learned a great deal about experimental design from him as well as how to navigate the research process as a Ph.D. student. The past two years that I’ve spent working with Martin have been immensely valuable.

Why are you interested in this subject?

When I first started on my research journey, I tried a ton of different things. My initial research focus was actually hardware research and then over the years I tried cryptography, privacy and machine learning. After much experimentation, I finally settled on social media because I realized how applicable it was to my personal life. As a social media user, I’m able to better understand the problems that users are facing because I fall into that demographic as well. With a social media focus, I can directly see the impact my research can have. 

Sometimes when we’re running experiments, I change my own feed to see how a different
algorithm affects my experience. Being able to directly measure and feel the impact of these changes is what keeps me motivated. Machine learning affects people in subtle ways, but this kind of work feels so tangible and immediate — it’s what keeps me excited about it.

What impact do you hope to make in the information field through your research/dissertation?

I hope that my research can, in some way, be applied by major social media platforms like
Instagram and Facebook to improve their algorithms and enhance the user experience. As a researcher, I’m faced with the challenge of balancing pro-social goals with business interests, and I aim for my work to contribute meaningfully to both.

What surprised you the most when digging into your research?

What surprised me most about research is its nonlinearity. While the final publication often appears organized and cohesive, the process behind it is much more chaotic. It involves exploring many directions, some of which lead nowhere, and doing a lot of work that ultimately never gets published. It’s a rewarding process, but one that can be messy and unpredictable at times.

What are your career goals once you graduate?
I hope to continue my research after graduation, whether through a career in academia or at an industrial research lab. If I pursue the academic path, I’d likely aim for a postdoctoral or faculty position at a research-focused university. In contrast, an industry role would come with a different set of responsibilities and incentives. My decision will also depend on how the research landscape and job market evolve over the next few years.