Unexpected Connections Emerge at CDS’ Weekly Data Science Lunch Seminar
This post is part of a series exploring CDS Seminars
In a field as vast and varied as data science, you never know where the next groundbreaking idea might come from. The Data Science Lunch Seminar Series at CDS provides a space for those serendipitous sparks of insight to fly.
Organized by CDS Faculty Fellows Yoav Wald and Elisha Cohen, the weekly series brings together data science students, faculty, and practitioners from across disciplines to share their latest work. “We strive to feature a mix of visiting scholars, faculty, and fellows from CDS, New York, and beyond,” Cohen explained. “It’s also important to us that all the incoming CDS Fellows have an opportunity to present their research.”
This thoughtful approach has yielded a diverse lineup of speakers, covering everything from linguistics to physics to the ever-popular large language models (LLMs). “The interdisciplinary nature of CDS naturally lends itself to a wide variety of topics,” Wald observed. “We’ve found that this diversity emerges organically based on the interests of our speakers and attendees.”
While the organizers don’t dictate topics, they do aim for broad appeal and intellectual engagement. Wald recalled a well-attended talk by Claudia Shi from Columbia on the morality of LLMs. “The topic was timely and thought-provoking,” he said. “It generated a lively discussion and exposed attendees to new perspectives at the intersection of technology and ethics.”
These thought-provoking discussions often spill over into impromptu meetings and collaborations, as attendees connect with speakers and each other to explore ideas in greater depth. After a recent talk by Princeton’s Benjamin Eisenbach, Wald noticed small groups of students and faculty engaged in animated conversation. “One of the great things about the lunch seminar series is the way it sparks organic interactions and collaborations,” he said. “We’ve seen firsthand how a casual conversation can plant the seeds for exciting interdisciplinary work.”
The series has been a fixture at CDS since the center’s early days, with the website archive stretching back to 2013. But Wald and Cohen are always looking for ways to enhance the experience for speakers and attendees alike. One area of focus is making sure attendees are aware of opportunities to schedule one-on-one meetings with speakers to discuss their work in greater depth.
They’re also exploring ways to expand access to the valuable content being shared each week. “We’ve begun recording the talks and have made them available upon request,” Cohen said.
On April 10, Julian Michael from NYU, spoke on “AI Alignment via Language Understanding,” and the final speaker of the season, Umang Bhatt, also from NYU, will speak on April 24. Bhatt researches trustworthy machine learning and human-machine collaboration.
As the Data Science Lunch Seminar Series evolves, it remains committed to fostering interdisciplinary connections and intellectual curiosity. With its engaging format and dedicated organizers, the series is poised to remain a vital part of the CDS community for years to come.
By Stephen Thomas
