ÌǹûÅɶÔ

The reliance on open-source software makes our infrastructure prone to the security vulnerabilities and safety failures of such software. Organizations often incur significant costs due to software maintenance, particularly for post-deployment bug fixes. Therefore, it is crucial to develop and implement robust analysis techniques during both the development and deployment stages.

The Trustworthy Systems Laboratory investigates whether systems can be proven to be trustworthy. Confidence in a system’s trustworthiness is achieved through reliable design, transparency, verification & validation, and various software analysis techniques such as static analysis, formal methods, and model checking.

Modern systems increasingly depend on event- and data-driven software components, making the assessment of their trustworthiness even more complex. These components often process large, dynamic datasets, integrate machine learning algorithms, and interact with diverse system environments. This introduces significant challenges in analyzing their behavior, predicting potential failures, and ensuring their security and reliability. Addressing these challenges requires innovative approaches capable of handling the dynamic and adaptive nature of both event and data-driven components. This includes enhancing existing methods for verification, validation, and analysis to account for data variability, algorithmic biases, and the evolving behavior of such systems.

Our vision is to facilitate proactive software analysis techniques to meet the global demand for design and systems that are not only reliable but also secure, robust, and resilient to failures.

Our laboratory investigates multiple dimensions of trustworthiness across software and systems, including human-machine interactions. We address critical aspects such as safety, functional correctness, predictability, security, and privacy. Additionally, we focus on traditional dependability attributes, including integrity, robustness, and reliability.

Services

• Analyze software artifacts to discover vulnerabilities/defects and provide mitigation.
• Utilize AI/ML techniques and build intelligent predictive models to optimize the trustworthy operation of various Cyber-Physical Systems (including robotics and automation).
• Automated program repair, i.e., finding and repairing software vulnerabilities and bugs on the fly.
• Consultancy to enhance confidence in system trustworthiness.

We focus on (but are not limited to) the following research lines: