Grigorios Tsoumakas

Papers

Total Citations

H-Index

About

Grigorios Tsoumakas is a leading researcher in explainable artificial intelligence and web security, whose work bridges cutting-edge machine learning with real-world cybersecurity challenges. His most influential contribution, "LioNets: Local Interpretation of Neural Networks through Penultimate Layer Decoding" (14 citations), addresses the critical need for transparency in AI systems powering smart homes, autonomous vehicles, and healthcare—offering a novel method to decode neural network decisions at the penultimate layer. Complementing this, Tsoumakas has pioneered semantic approaches to web robot detection, with papers such as "Web Robot Detection: A Semantic Approach" (13 citations) and "Content-aware web robot detection" (10 citations) tackling the growing threat of malicious bots that now constitute over half of global web traffic. His work extends to academic publishing ("Web Robot Detection in Academic Publishing," 2 citations), where he exposes how bots distort analytics and metrics. Across these contributions, Tsoumakas demonstrates a rare ability to address both the interpretability of complex neural models and the practical security challenges of the modern web, making his research essential for students and practitioners working at the intersection of trustworthy AI and cybersecurity.