Summer School on Strings & Privacy

Dates

June 13th and 14th, 2026

Speakers

Teresa Anna Steiner, Solon Pissis

Topics

Introduction to Differential Privacy
Differentially Private Pattern Matching
Combinatorial Sanitization of Strings

Location

DTU, Building 303A Room 045

How to get there

From the city center, the easiest way to get to DTU is to take line 150S from Nørreport station to Rævehøjvej, DTU (Helsingørmotorvejen). It takes about 45 min from Nørreport to the venue. We recommend using the app Rejsebillet to buy your ticket and use Rejseplanen (app or .dk) to find your journey.

Price

Free. Lunch & social event included for PhD students and Postdocs.

Registration deadline

Please register by May 31st. Registration link can be found at the bottom of this page.

Overview

In the stringology community, we are usually concerned about analyzing strings efficiently. However, when the data we analyze contains sensitive data (for example, DNA sequences or location tracking), another important question is how we can compute accurate information about the strings while preserving privacy.

The goal of this summer school is to introduce the exciting new research fields of privacy preserving string analysis and string sanitization. It will introduce different definitions of what it means to sanitize strings and give tools for designing privacy preserving algorithms for strings. We will discuss the latest research results in this area and open problems.

Programme

Saturday June 13th

Time: 10-12h and 13-16h (lunch 12h-13h)

Differential Privacy & Strings (Teresa Anna Steiner)

Differential privacy is considered the gold standard for privacy in data analysis. It gives rigorous, probabilistic guarantees about the contribution of a single user being “hidden” in the output of an algorithm. The definition of differential privacy naturally comes at a loss of accuracy. In differential privacy research, the goal is to design algorithms which are as accurate as possible under a fixed privacy guarantee.

In the first part of the lecture, we introduce the concept of differential privacy. We will see different tools for designing differentially private algorithms and how to analyze them.

In the second part of the lecture, we will see how the concept of differential privacy can be applied to strings. We will discuss several possible models for differential privacy for strings, what they promise in terms of privacy, and what challenges they pose for designing algorithms. We will see state-of-the-art algorithms and data structures for pattern matching variants under differential privacy.

Sunday June 14th

Time: 10-13h

Combinatorial Sanitization of Strings (Solon Pissis)

String data are often disseminated to support applications such as location-based services or DNA sequence analysis. This dissemination, however, may expose sensitive patterns that encode confidential knowledge (e.g., trips to mental health clinics from a string representing a user's location history). In the first part of this lecture, we will discuss the general problem of sanitizing a string by concealing occurrences of sensitive patterns, while maintaining data utility, in settings relevant to common string processing tasks.

In the second part, we will turn our attention to data structures. Real-world applications rarely access raw data directly; instead, they rely on data structures constructed from them. Ensuring that these structures do not allow the original data to be reconstructed is therefore essential for privacy. Our goal is to introduce a paradigm for constructing data structures that prevent reconstruction of their underlying input, while still supporting many useful queries. To this end, we lift the k-anonymity privacy principle to the data structure setting.

A unifying theme of both parts is that they give rise to interesting combinatorial problems. We will discuss exact and optimal algorithms to address them, as well as open problems for future research.

Time: 13h: Lunch

Time: Afternoon: Social event