Jamie Lohoff
Papers
1
Total Citations
2
H-Index
1
About
Jamie Lohoff is a rising researcher at the forefront of computational efficiency, whose work bridges the critical gap between machine learning and scientific computing. Lohoff’s primary research focuses on optimizing automatic differentiation (AD)—the backbone of modern machine learning and simulation—by leveraging deep reinforcement learning to reduce the computational and memory costs of Jacobian computations. In their seminal 2024 paper, "Optimizing Automatic Differentiation with Deep Reinforcement Learning," Lohoff demonstrated how even modest savings in AD operations can translate into massive energy and time reductions across fields like computational fluid dynamics, robotics, and finance. Though early in its trajectory, this work has already garnered attention for its novel cross-disciplinary approach. Lohoff’s contributions are particularly impactful because they address a universal bottleneck: the inefficiency of gradient calculations in large-scale systems. By framing AD optimization as a reinforcement learning problem, Lohoff opens the door to self-improving numerical software. Their research promises to accelerate everything from neural network training to climate simulations, marking Lohoff as a key innovator in the quest for sustainable, high-performance computing.
Research Focus
Key Achievements
Top Papers
- 1Optimizing Automatic Differentiation with Deep Reinforcement Learning2 citations · 2024