Orr Krupnik

Technion – Israel Institute of Technology

Papers

Total Citations

H-Index

About

Orr Krupnik is a researcher specializing in multi-agent reinforcement learning and model-based approaches to autonomous decision-making. His work sits at the intersection of robotics, machine learning, and multi-agent systems, with a particular focus on solving challenging continuous control problems where multiple agents must interact within shared environments — a domain with significant real-world relevance for collaborative and competitive robotics. Krupnik's most notable contribution, "Multi-Agent Reinforcement Learning with Multi-Step Generative Models" (2019), addresses one of the fundamental challenges in model-based reinforcement learning: the accumulation of prediction errors over time. In multi-agent, high-fidelity continuous control settings, this problem is especially pronounced, and his research proposes generative modeling strategies to mitigate these compounding inaccuracies. The work has garnered 7 citations, reflecting early but meaningful traction within the specialized community of robotics and multi-agent systems researchers. While still an emerging voice in his field, Krupnik's research tackles problems of genuine practical importance — particularly as autonomous robots increasingly operate alongside other agents in complex, dynamic environments. Students interested in the frontier of model-based multi-agent learning will find his work a valuable entry point into these technical challenges.