Speaker
Description
Corruption in public procurement, financial transactions, and administrative processes presents substantial threats to national and international security, compromising integrity in defense logistics, resource distribution, and governance frameworks. Artificial intelligence, especially machine learning models for finding unusual patterns, spotting fraud, and predicting risk, has become a powerful tool in the fight against corruption. But big AI models often have trouble being used in secure, resource-limited settings (like edge devices used for military or public sector monitoring). These problems include high computational needs, susceptibility to adversarial attacks, and limited interpretability, all of which can make them less trustworthy and secure.
This study examines model pruning techniques, including structured and unstructured pruning, as well as magnitude-based and lottery ticket hypothesis-inspired approaches, to create compact and efficient anti-corruption AI models with minimal accuracy degradation. These pruned models run faster, use much less memory, and can be used on edge devices or in secure environments with limited resources. They also do a great job of finding key corruption signals like bid-rigging, suspiciously short tender deadlines, single-bid wins, or irregular fund flows.
The proposed approach offers a pathway for integrating robust, lightweight anti-corruption AI into defense and security technologies, supporting decision support systems and infrastructure protection. This work bridges artificial intelligence advancements with secure technology development for enhanced societal integrity and resilience.