The KDD Lab program in secure and trustworthy computing focuses on methods for assessing the soundness, safety, and security of sequential problem solving systems. Methods for doing so include:

• adversarial robustness - cf. Behzadan (2019)
• explainable artificial intelligence (XAI) - for inference in computer vision, for reinforcement learning (XRL)
• grounding of language models - question generation cf. Lamba (2021)
• document understanding - information extraction cf. Yang (2022), graph representation and geometric machine learning cf. Bose (2022)

Applications of this work include:

• Cyber threat intelligence (CTI)
• Disinformation monitoring in social media and networks

The KSU KDD Lab is an affiliate member of the K-State Center for Cybersecurity and Trustworthy Systems (K-CaTS).

Last updated by bhsu on Mar 24, 2024