Donald Perlis
University of Maryland, College Park, Williams (United States)
Papers
5
Total Citations
26
H-Index
3
About
Donald Perlis is a pioneering researcher in artificial intelligence, with a career spanning foundational work in commonsense reasoning and temporal logic to cutting-edge deep learning for robotics. His key research areas include the frame problem, time-sensitive reasoning, and embodied cognition. Perlis is best known for his seminal 1991 paper, "Stop the world—I want to think" (11 citations), which fundamentally reframed the frame problem by arguing that rational agents must explicitly manage the tension between deliberation and the passage of time. This work remains a cornerstone for researchers tackling real-time AI systems. More recently, Perlis has made significant contributions to autonomous robotics, particularly in size, weight, and power (SWaP)-constrained platforms. He led the development of novel deep neural network architectures, such as the Multi-Hypothesis DeepEfference (MHDE) network (2018, 4 citations), which fuses vision and motion data for robust state estimation. His work on "DeepEfference" (2017, 3 citations) draws inspiration from biological visual constancy to improve robotic localization. Perlis’s research uniquely bridges high-level reasoning about time and self-awareness with low-level sensorimotor learning, demonstrating a sustained impact on both theoretical AI and practical robotic perception.
Research Focus
Key Achievements
Top Papers
- 1Stop the world—I want to think11 citations · 1991
- 2
- 3
- 4Reasoning with Grounded Self-Symbols for Human-Robot Interaction.3 citations · 2016
- 5