A research paper co‑authored by Inge Li Gørtz and Philip Bille has received the SIGMOD Research Highlight Award 2026, an international award presented by ACM SIGMOD to research that “exemplify core database research. In particular, these projects address an important problem, represent a definitive milestone in solving the problem, and have the potential of significant impact on the research field of databases and data management,” writes SIGMOD.
The award‑winning paper, “Differentially Private Substring and Document Counting”, was published in the Proceedings of the Symposium on Principles of Database Systems (PODS 2025), where it also received the Best Newcomer Award.
The SIGMOD Research Highlight Award is presented annually to a very limited number of papers that stand out by:
- addressing fundamental problems in data processing,
- delivering technically deep and broadly applicable solutions, and
- having the potential for long‑lasting impact on both theory and practice.
Efficient text analysis with strong privacy guarantees
The paper, authored by (in alphabetical order): Giulia Bernardini (University of Milan), Philip Bille (DTU Compute), Inge Li Gørtz (DTU Compute), and Teresa Anna Steiner (University of Southern Denmark), addresses a central question in modern data analysis: How can large collections of text be analysed without compromising individual privacy?
In the paper, the researchers introduce a differentially private data structure for indexing text collections, enabling counting queries for arbitrary patterns – for example, how often a given word or phrase occurs – with almost optimal accuracy, while ensuring that each individual document is protected by formal, mathematical privacy guarantees.
The results show that the method is particularly well-suited for datasets consisting of many short or medium‑length texts, a common scenario in practice, such as in digital communication data and log files.
Combining string algorithms and differential privacy in a new way
The improvements build on new combinatorial observations about the structure of the problem and on a novel combination of techniques from string algorithms and differential privacy under continual observation.
The resulting data structure is not only theoretically strong, but also conceptually simple, space‑efficient, and fast to query.
Former DTU PhD highlighted as expert in privacy
A central contributor to the work is Teresa Anna Steiner, who is internationally recognised as an expert in differential privacy and privacy‑preserving algorithms. She is a former PhD student at DTU, supervised by Inge Li Gørtz and Philip Bille, and is currently employed as an assistant professor at the University of Southern Denmark.
Her research focuses on developing algorithms and data structures with strong privacy guarantees, particularly at the intersection of privacy and string algorithms – the very area in which this work represents a significant breakthrough.
“We are very honoured to receive the award,” says Inge Li Gørtz, Professor at DTU Compute.
“It is particularly gratifying to see how the expertise in differential privacy that Teresa developed during her postdoctoral work now plays a central role in research that is being highlighted internationally for its impact.”
