The Horizon Europe project CHESS – Cyber-security Excellence Hub in Estonia and South Moravia co-sponsored the 23rd Estonian Summer School on Computer and Systems Science (ESSCaSS 2026), held from 18–21 August in the coastal village of Käsmu, Estonia.
Organized by the University of Tartu Institute of Computer Science, ESSCaSS is an annual international event for PhD and MSc students, postdoctoral researchers, faculty members, and other specialists. Its interdisciplinary program spans computer science, systems science, software engineering, artificial intelligence, bioinformatics, and related fields.
CHESS contributed to the summer school through invited lectures, research presentations, and interactive sessions, reflecting the CHESS project’s commitment to strengthening cybersecurity research, innovation, and knowledge exchange across Estonia, South Moravia, and the wider European community.
From certification processes to real-world security
One of the program’s cybersecurity highlights was a two-part presentation by Professor Václav “Vashek” Matyáš of Masaryk University.
In “Process, Proof, and Practice: A Deep Dive into Common Criteria Security Certification,” Matyáš examined the Common Criteria certification ecosystem. His second topic, “The Reality of Certified Security: Mining the CC and FIPS Ecosystems with Sec-Certs,” turned to the practical realities of certified products. Matyáš presented research based on large-scale automated analyses of certification documents and demonstrated how the open-source sec-certs toolkit can link known vulnerabilities to certificates and reveal dependencies among certified products.
Exploring the future of program verification
CHESS also organized a 90-minute verification session led by Karoliine Holter and Vesal Vojdani. The session offered an accessible introduction to program verification with concrete examples before moving on to emerging AI-assisted approaches such as “vibe proving.”
Participants explored why program correctness matters, how verification relates to programming-language research, and the trade-offs among automation, precision, and scalability. The session also considered how formal verification and sound static analysis could enhance the reliability of AI-driven software development.
TWINE brings explainable AI to industrial anomaly detection
Mubashar Iqbal presented a poster based on the recently published research paper “TWINE: ISO 23247-compliant digital twin and explainable AI framework for anomaly detection.” The article appears in Elsevier’s Internet of Things journal.
TWINE combines digital twin technology with explainable artificial intelligence to enable real-time anomaly detection in industrial robotic arm operations. Designed in accordance with the ISO 23247 standard, the framework provides a modular and interoperable connection between physical manufacturing equipment and its digital representation.
By combining supervised machine learning with SHAP-based explanations, TWINE not only detects abnormal behavior but also helps operators understand which factors influenced each result. This transparency is particularly important in industrial environments, where trustworthy and interpretable decisions can help reduce downtime, improve product quality, and support safer operations.
