Tianji Cong receives Rackham Predoctoral Fellowship

CSE PhD student Tianji Cong has been selected to receive a 2024-25 Rackham Predoctoral Fellowship. The fellowship will support his ongoing research, which focuses on the application of representation learning techniques to extract meaningful information from relational and tabular data.

Awarded to a select group of doctoral candidates each year, the Rackham Predoctoral Fellowship recognizes students who have demonstrated outstanding accomplishments and/or progress toward completion of their dissertation. Awardees are also selected based on the results of their work, such as published papers or conference presentations, as well as excellence in teaching or other professional service.

Artificial intelligence (AI) has seen rapid advances in recent years, and there is a great desire today to use AI techniques in data management systems to strengthen how data is stored, organized, and accessed.  While much progress has been made in this regard to individual databases, there is little work as yet on repositories comprising thousands, or even millions, of databases.

Cong’s research looks at the design and application of AI techniques to enrich and search for tabular data in a repository. His work involves the use of representation learning, an AI technique broadly used in areas spanning natural language processing to computer vision, to comprehend and analyze tabular and relational data. His research has yielded new methods of data discovery and database management for these types of data, which have had a broad impact across academia and industry.

“The creative nature of Tianji’s work, on a contemporary topic of great importance, is in itself enough to be praiseworthy,” said Prof. H.V. Jagadish, Cong’s advisor and nominator, “But what I find truly remarkable is Tianji’s span from conceptual advances in academic papers to real implementations in deployed systems.”

An example of this is the deployment of Cong’s prototypes by Sigma Computing, a cloud analytics startup, as well as by the Research Data Ecosystem at the Inter-university Consortium for Political and Social Research (ICPSR), the world’s largest social science data archive. His work has also resulted in several publications at prestigious conferences, including VLDB, CIDR, and NeurIPS. 

With the support of the Rackham Predoctoral Fellowship, Cong plans to continue his development of representation learning techniques for data management, expanding the scope of AI applications in data science and transforming the data structures landscape for the better.