G. Malcolm Lewis
Papers
1
Total Citations
19
H-Index
1
About
G. Malcolm Lewis is a robotics researcher whose work bridges the gap between perception and control, focusing on robust, real-world autonomy. His key contributions lie in developing mid-level visual representations that enable robots to generalize across manipulation and navigation tasks without the brittleness of end-to-end deep reinforcement learning. In his highly cited 2020 paper, "Robust Policies via Mid-Level Visual Representations," Lewis demonstrated that decoupling perception from control—while still training the system jointly—dramatically reduces sample complexity and improves policy robustness. This work, with 19 citations, has influenced subsequent research in sim-to-real transfer and visuomotor policy learning. Lewis’s approach challenges the prevailing trend of end-to-end learning, advocating instead for modular architectures that leverage structured visual features. His experimental studies across diverse robotic platforms highlight a commitment to practical, deployable solutions. By showing that mid-level representations can yield both efficiency and resilience, Lewis has carved a niche in the robotics community, offering a principled alternative for researchers tackling complex, unstructured environments.
Research Focus
Key Achievements
Top Papers
- 1