Alexander Trott

Papers

Total Citations

H-Index

About

Alexander Trott is a researcher whose work lies at the intersection of deep reinforcement learning and multi-agent systems, with a focus on developing algorithms that can learn effectively in complex, sparse-reward environments. His most notable contribution is the introduction of "Competitive Experience Replay," a novel framework that leverages competitive self-play to generate a natural curriculum of increasingly challenging tasks. This approach addresses a critical bottleneck in reinforcement learning: the need for carefully shaped reward functions. By enabling agents to learn from their own competitive interactions, Trott's work reduces the reliance on dense, hand-crafted rewards, making RL more scalable and robust. While his 2019 paper on this topic has garnered early citations, its impact is already recognized as foundational for advancing multi-agent training paradigms. Trott's research continues to push the boundaries of how autonomous systems can learn from sparse feedback, with potential applications in robotics, game AI, and autonomous driving. His work exemplifies a move toward more sample-efficient and generalizable reinforcement learning methods.