Miljan Martic
Papers
1
Total Citations
508
H-Index
1
About
Miljan Martic is a leading researcher in artificial intelligence, best known for his pioneering work at the intersection of deep reinforcement learning and human-computer interaction. His most influential contribution, the 2017 paper "Deep reinforcement learning from human preferences," has garnered over 500 citations and introduced a paradigm-shifting method for training AI systems using non-expert human feedback. Rather than requiring complex reward engineering, Martic demonstrated how agents could learn sophisticated behaviors by simply comparing pairs of trajectory segments based on human preferences—a breakthrough that made RL systems far more practical for real-world deployment. This work has become foundational in the field of AI alignment and preference-based learning, enabling safer and more controllable autonomous systems. Martic's research addresses one of the central challenges in modern AI: how to communicate complex, nuanced goals to learning systems without explicit programming. His contributions have influenced subsequent work in reinforcement learning from human feedback (RLHF), a technique now widely used in large language models and robotics, cementing his role as a key figure in making advanced AI systems more responsive to human values and intentions.
Research Focus
Key Achievements
Top Papers
- 1Deep reinforcement learning from human preferences508 citations · 2017