Sixteen students from around the U.S. gathered together on the Rice University campus this summer for a data science research internship alongside high-profile faculty mentors and their doctoral students.
Rice’s Department of Computer Science launched the 10-week data science summer program for undergraduates on May 24th. The Research Experience for Undergraduates, Data Science (REU DS) program was sponsored by Google and Nancy Dunlap with additional funding from the Department of Computer Science. The main goal was to provide students with unprecedented one-on-one access to faculty and doctoral students to conduct cutting-edge research.
REU DS also provided the interns interaction with Google employees to learn about their data science research. Students regularly attended faculty and expert data science seminars as well as the presentations from their peers on other data science research happening in the program.
Participating undergraduate students came from universities all over the US, including MIT, Johns Hopkins and Cornell as well as from Texas universities such as the University of Houston, Texas State University, St. Thomas University and Rice University.
“The students were extremely bright, curious, driven, ready-to-learn, and fun-loving,” says Sanjoy Paul, director of the REU Data Science program. “Because of their readiness to learn and deep curiosity about data science—together with the sincere coaching provided by the Rice mentors and co-advisors from Google—the interns had an incredible exposure to a wide range of R&D topics and an awesome R&D experience.”
Yawen Ding, an undergraduate from Cornell, worked on Density Sketches and Coresets research with associate professor Anshumali Shrivastava and doctoral student Aditya Desai. She says, “It was nice having one-on-one time with the professors and PhD students that I work with. This is something I don't get often in a regular semester. I'm happy that they had lots of projects available to me that fit my interests.”
Chima Adiole, an undergraduate at Rice, worked alongside doctoral student Charles Jiang and faculty member Chris Jermaine, chair of the Department of Computer Science as well as the former director of Rice’s Data Science Initiative. She’s researching program synthesis.
“The project I am currently working on has to do with using deep learning techniques to understand and process structured data,” Adiole says. “I have been able to learn a lot about applying various techniques within data science to approach complex problems, and the importance of interdisciplinary collaboration for advancement in research. My PI and PhD student mentor have been extremely supportive throughout the whole process, and I am excited to continue research within this field.”
This summer’s projects included research into a very broad range of data science and machine learning topics including designing and implementing new machine learning algorithms, applying machine learning algorithms to solve specific problems and developing methods to manage huge data sets. Research touched on a range of fields including finance, civic tech and healthcare.
Student researchers presented findings from their summer research during a poster session on Friday, July 30th on the Rice campus. Following the presentations, program organizers and Google Vice President and Engineering Fellow Parthasarathy Ranganathan also addressed the crowd.
The end of the summer culminated with a Data Science Project Competition among the students. The winners, announced during the closing ceremony, included Justin Lumpkin in first place for his Empirical Analysis of Redundant Linear Layers in Deep Learning research under the guidance of Rice Computer Science Noah Harding Assistant Professor Anastasios Kyrillidis and computer science doctoral students J. Lyle Kim and Chen Dun.
Student Tessa Cannon placed second with her research on Improving Harris County Census Data with Density Mapping and Machine Learning. She worked under faculty mentor and Computer Science Associate Progefssor Anshumali Shrivastava and his graduate student mentor Zhaozhuo Xu.
Finally, John Poole and Cruz Barnum tied for third place. Barnum’s research focused on Providing Bounds for k-PCA, also under the guidance of Kyrillidis, Kim and Dun. Poole’s research, titled Neural Networks for Feature-Selection in Operator Induced Structures, was conducted under the supervision of faculty member Meng Li, Noah Harding Assistant Professor of Statistics and her doctoral student Huiming Lin.
For information on applying to the REU DS program, please visit cs.rice.edu/reu.