News
AI against money laundering
In a new research project, researchers at Fraunhofer SIT are developing a solution for the automated detection of illegal financial flows - including testable minimum requirements for the use of artificial intelligence in sensitive areas.
read more
Accepted papers on ACM SIGMOD 2023
Three papers involving ATHENE researchers were accepted at the A*-ranked 2023 ACM SIGMOD/PODS Conference. The annual conference is a leading international forum for database researchers, practitioners, developers, and users to explore innovative ideas and results and to share techniques, tools, and experiences. The conference offers a comprehensive technical program with presentations from research and industry, tutorials, demos and workshops.
read more
Successful submission to PLDI 2023
ATHENE researchers are placing a paper at the highly ranked 44th ACM SIGPLAN Conference on Programming Language Design and Implementation (PLDI 2023). The conference is the leading forum in the field of programming language and programming systems research.
read more
Successful submissions at ICSE 2023
Two papers written in collaboration with ATHENE researchers were accepted at the 45th IEEE/ACM International Conference on Software Engineering. ICSE, the IEEE/ACM International Conference on Software Engineering, is the leading conference on software engineering. Since 1975, ICSE has provided a forum where researchers, practitioners, and educators come together to present and discuss the latest innovations, trends, experiences, and issues in the field of software engineering.
read more
ChatGPT or human author?
ATHENE researchers at Fraunhofer SIT are researching ways to help recognize texts created with ChatGPT. Among other things, the text forensics experts are working with a self-developed method for authorship verification, COAV: Originally, it was used to detect plagiarism in scientific papers, for example. Since COAV compares texts on a stylistic basis, this method can also be used to identify a specific "author," namely ChatGPT. This is used to calculate the distances between texts using similarities of text modules and typical consecutive letter strings: Is the text closer to GPT or closer to a human?
read more
Whitepaper on "Active Cyber Defense"
As a result of the Russian attack on Ukraine, the discussion about active cyber defense has also flared up again. Politicians are calling for improved capabilities. In their white paper "Active Cyber Defense" our CEO Prof. Michael Waidner and our cybersecurity expert Prof. Haya Shulman provide concrete examples of the technical options available for improving cyber defense in Germany.
read more
ATHENE researchers present document protection with colorful barcode
At this year's it-sa in Nuremberg, Fraunhofe SIT is showcasing DocSeal, a new solution for protecting against document forgery that enables companies and public authorities to quickly and easily add anti-counterfeiting protection to digital and paper documents. For this purpose, a colorful barcode (JAB Code) is printed on the document, which records important document contents and their placement in the document in a tamper-proof manner. An app can then be used to check document authenticity and automatically detect tampering.
read more
RPKI is insecure - Mechanism for Internet security broken
ATHENE has found a way to break one of the basic mechanisms used to secure Internet traffic. The mechanism, called RPKI, is actually designed to prevent cybercriminals or government attackers from diverting traffic on the Internet. Such redirections are surprisingly common on the Internet, e.g., for espionage or through misconfigurations. The ATHENE scientist team of Prof. Dr. Haya Shulman showed that attackers can completely bypass the security mechanism without the affected network operators being able to detect this. According to analyses by the ATHENE team, popular implementations of RPKI worldwide were vulnerable by early 2021. The team informed the manufacturers, and now presented the findings to the international expert public.
read more
Client Side Scanning and Deep Perceptual Hashing Vulnerabilities
ATHENE scientists at TU Darmstadt have identified significant vulnerabilities and manipulation possibilities in client-side scanning and deep perceptual hashing. The process came into focus when Apple introduced "NeuralHash" in 2021, a new approach to detecting child abuse imagery, but withdrew the introduction after massive criticism. The research results of the scientists now prove the dangers of client-side scanning methods for users.
read more
Fragile protection of our communications via submarine cables
ATHENE researchers study states' vulnerability to submarine cable failures
Today, we take it for granted that we can call up a website, stream a movie or be active in social networks within seconds. Many people are often unaware that the data transfer takes place via thousands of kilometers of cable laid at the bottom of the ocean. Today, around 98 percent of international Internet traffic is handled via undersea communication cables. Coastal and island states are highly dependent on this physical infrastructure to provide Internet connections. However, although an annual average of about 100 submarine cable failures of human or natural origin occur, there is currently no global analysis that assesses the vulnerability of individual states to failures on a global scale.
ATHENE scientists Jonas Franken, Thomas Reinhold and Prof. Christian Reuter from the Chair of Science and Technology for Peace and Security (PEASEC) at TU Darmstadt have tackled this issue.
