Jakob Foerster

Papers

Total Citations

H-Index

About

Jakob Foerster is a leading researcher in multi-agent reinforcement learning (MARL) and emergent communication, whose work is shaping how AI systems learn to cooperate and communicate. His major contributions center on developing theoretical frameworks and algorithms that enable multiple agents to solve complex coordination problems, particularly in partially observable environments. Foerster’s research addresses fundamental challenges in Dec-POMDPs—settings where agents must act based on limited information—by introducing centralized models and exploration policies that dramatically improve learning efficiency. His work on quasi-equivalence discovery for zero-shot emergent communication (2021, 6 citations) has opened new pathways for agents to develop effective communication protocols without prior training, a critical step toward scalable multi-agent systems. Additionally, his application of reinforcement learning to soft robotics (2024, 5 citations) demonstrates the practical impact of his methods, tackling the nonlinear dynamics of compliant manipulators. With a citation count that reflects growing influence, Foerster is recognized for bridging theory and application, making him a pivotal figure in advancing cooperative AI for real-world challenges like robotic swarms and autonomous teams.