Aditya Deshpande

Indian Institute of Technology Hyderabad, University of the Cumberlands

Papers

Total Citations

H-Index

About

Aditya Deshpande’s research spans computer vision and cybersecurity, with a focus on geometric scene understanding and the safe deployment of AI agents. His early work tackled the fundamental challenge of detecting multiple planes from image pairs, a problem critical for 3D reconstruction and robotics. By building on methods like J-Linkage, he contributed to domain-independent approaches for multiple structure detection, laying groundwork for robust visual perception. More recently, Deshpande has pivoted to the emerging frontier of AI safety, addressing the cybersecurity risks of Large Language Model-driven robotic agents conducting online transactions. His 2024 study proposes enforcing robust constraints to prevent data breaches and transaction fraud, a timely contribution as autonomous systems become more integrated into e-commerce. While his citation counts are modest—2 and 1 respectively—the breadth of his work demonstrates adaptability and foresight, tackling both foundational vision problems and cutting-edge AI security. For students and researchers, Deshpande’s trajectory offers a compelling example of how technical expertise can evolve to meet new challenges, from geometric inference to safeguarding the next generation of autonomous agents.