Research Focus Class on Medical Data Privacy

Our Research Focus Classes (RFC) are a special kind of lecture: they are more interactive and research-oriented than typical lectures. Students participating in an RFC should be aware that they are not only getting in touch with real research but also have to expect doing independent work. In the past, results from our RFCs have lead to publications at high-ranking scientific venues and many participants continued to pursue their sparked passion for research within a PhD.

Please note: If you are interested in participating in the upcoming winter term, please drop us an email (rfc(at)comsys.rwth-aachen.de) and attach a few sentences stating your motivation of taking this course besides a current transcript of records.

Motivation

In the contemporary landscape of medical research, data is the lifeblood propelling groundbreaking advancements. Modern medicine increasingly relies on vast datasets; the depth and breadth of which fuel our understanding of complex diseases, genetic variations, and patient outcomes. The surge in data-driven approaches, particularly the rise of machine learning and artificial intelligence, has already offered unparalleled insights: With larger quantities of data, machine learning models refine their predictions, making diagnostic tools more accurate and treatments more personalized than ever before. However, with this influx of sensitive information, privacy becomes paramount. Ensuring that individuals can trust the medical community with their intricate health data is not just an ethical imperative but also a foundational element that bolsters research. This trust motivates greater participation in studies, leading to richer datasets and setting the stage for transformative medical breakthroughs.

Balancing the immense potential of data with the inviolable right to privacy presents a challenge. Innovative solutions that prioritize data privacy foster trust, encourage collaboration, and facilitate wide-reaching studies that can revolutionize the medical field. By delving into state-of-the-art encryption methods, differential privacy, distributed analytics, and FAIR data management methods, researchers can harness the power of medical data while rigorously protecting individual rights. Join our RFC to explore this intersection of medicine and privacy, and equip yourself with the expertise to support future medical innovation.

 

Organizational Information

  • SWS: V3/Ü2, ECTS: 6
  • Organizers: this course is offered in cooperation between COMSYS and the UKA
  • Study programs: Master Informatik (Software und Kommunikation), Master Software Systems Engineering (Communication), Master Media Informatics, Master Data Science, Bachelor Informatik (nach Absprache as Vorzugsfach)
  • Language: English
  • Start: TBD, KW21
  • Lecture slot: Doodled among participants
  • Location: COMSYS, Building E3, Ahornstr. 55
  • Due to limited capacity, prior registration is required!
  • Registration and questions: please send an email to rfc [[at]] comsys.rwth-aachen.de

Involved Partners

We are grateful for the support of our partners with a background in medince, primarily from the University Hospital Aachen (Uniklinik Aachen), who allow us to conduct interdisciplinary research in this RFC on Medical Data Privacy.

Content

In this RFC, you will get in touch with current security, privacy, or data managment problems of medical (big) data, state-of-the-art security solutions and missing paradigms for security in (envisioned) future medical research. Depending on your interests, you will dive deeper into topics such as:

  • Discovering and Exploring Data in Distributed Data Silos
    • On the way to a medical big data landscape, privacy regulation, patient consent and legacy data management systems have created large (and mostly isolated) data silos that nowadays limit the efficacy of data driven research, such as via big learning. An urgent challenge thus is to break up these silos in a privacy-preserving and regulation-compliant way to enable data access, e.g., by filtering for specific symptoms relevent to studying a disease.
  • PETS to process and analyze sensitive data
    • Recent cryptographic advancements, such as the feasibility of Fully Homomorphic Encryption, hardware-based approaches for confidential computing, and computational architectures, such as distributed analytics nowadays enable us to study privacy-sensitive medical data. Here, you will learn how to apply and extend these techniques to medical research questions, including training sophisticated models or utilize complex data, such as MRI images.
  • Sharing medical datasets in privacy-preserving ways and retrospectively analyzing anonymity
    • Public data sources, such as BioBanks, but also datasets released as part of prior medical or data scientific competitions significantly contribute to broad scale medical research. At the same time, publishing medical data regularly is a huge privacy challenge as data needs to be deidentified and anonymized. Here, you will learn about tools to deidentify data, as well as methods to check their efficiency.

Structure

The RFCs are research-oriented courses following an interactive schema. To do so, we give a short introductory lecture about the topic (5-6 lecture slots). This lecture phase is accompanied by small practical homework tasks to familiarize yourself with the topic. Afterward in the analysis phase, you identify your own interesting research projects within the scope of medical data privacy (if necessary, we will help you with this) and prepare a short presentation of the motivation behind your idea and how you intend to tackle it. In the remainder of the class you then should realize your idea in the form of a mini project. The RFC concludes with a poster session where you will present your results. This year, you will also have the opportunity to instead draft a paper, with the goal of submitting it to a workshop or conference.

Overall Schedule

  1. Lecture Phase: 2-3 weeks
    • Getting up to speed!
    • Learn about concepts, approaches, tools, examples ...
    • Hands-on supplementary homework tasks to practice what you learned
  2. Concept Phase: 4 weeks
    • Get to know our medical experts and their privacy problems
    • Based on that: Develop your own resarch idea
    • Present and discuss your idea with the other participants
  3. Mini Project: up to 12 weeks
    • Get your hands dirty! Perform work on your research idea
    • Short presentation/demo at the end
    • Interesting/fruitful projects may even result in a scientific publication

Prerequisites

This class will serve as an introduction on how to conduct research in the area of data security, privacy and management. Besides learning about the actual topics, taking this course is an ideal preparation for doing a master thesis in one of our groups and later on pursue a career in research. You should have prior knowledge at least in the basics of data security and an interest in doing independent research. A background in medicine is not needed. As this course is supposed to be highly interactive, seats are limited. Should we receive more registrations than seats available, we will select students based on their qualification for this course.

- Impressum | Datenschutz -