Marcus A. Pereira

Papers

Total Citations

H-Index

About

Marcus A. Pereira is a researcher working at the intersection of stochastic optimal control theory, applied mathematics, and machine learning, with a particular focus on developing principled frameworks for decision-making under uncertainty. His most influential work, "Learning Deep Stochastic Optimal Control Policies using Forward-Backward SDEs" (2019, 25 citations), introduced a groundbreaking methodology that bridges nonlinear partial differential equations and deep learning to solve complex control problems — a contribution that has resonated strongly across robotics, control theory, and computational mathematics communities. Building on this foundation, Pereira extended his approach to handle more challenging control settings, including systems with control multiplicative noise, demonstrating the versatility of deep recurrent neural network architectures in tackling fully nonlinear Hamilton-Jacobi-Bellman equations. Perhaps most impressively, his work transcends traditional engineering boundaries: his 2020 paper on neuromechanical systems applies stochastic optimal control as a computational analogy for understanding how the nervous system governs tendon-driven bodies, reflecting his ambition to connect mathematical rigor with biological insight. Across his body of work, Pereira has established himself as a creative researcher pushing the boundaries of what modern machine learning can achieve in principled control and decision-making.