Logan Engstrom

Papers

Total Citations

H-Index

About

Logan Engstrom is a prominent researcher in computer vision and machine learning, best known for his work on adversarial robustness and the practical design of vision systems. His highly cited paper, "Unadversarial Examples: Designing Objects for Robust Vision" (2020, 25 citations), introduces a groundbreaking framework that flips the traditional adversarial paradigm: instead of defending against attacks, Engstrom shows how to proactively design objects to improve model performance and robustness. This work highlights his focus on understanding and leveraging model sensitivities to create more reliable AI systems. Engstrom’s broader contributions span adversarial machine learning, where he has developed tools and benchmarks that have become foundational in the field. His research has been recognized for its impact on both theory and real-world deployment, earning him a reputation for bridging the gap between robustness research and practical application. With a citation count reflecting his influence, Engstrom continues to shape how vision models are built, tested, and hardened against failure, making his work essential reading for students and researchers interested in the future of trustworthy AI.