Franziska Herbert
Papers
1
Total Citations
12
H-Index
1
About
Franziska Herbert is a leading researcher in interactive reinforcement learning (IRL), with a focus on integrating multimodal human feedback to accelerate machine learning. Her most-cited work, "Interactive Reinforcement Learning With Bayesian Fusion of Multimodal Advice" (2022, 12 citations), pioneers the fusion of speech and gesture inputs into IRL systems, enabling more intuitive and efficient human-robot collaboration. By developing Bayesian frameworks that reconcile diverse advisory signals, Herbert has addressed a critical bottleneck in RL: the slow convergence of learning algorithms. Her contributions are foundational to creating AI agents that can learn rapidly from natural, multimodal human instruction—a key step toward deployable, interactive intelligent systems. Beyond this flagship paper, Herbert’s research continues to shape how machines interpret and act upon complex, real-time human guidance, with implications for assistive robotics, autonomous systems, and human-AI teaming. Her work is increasingly cited in both machine learning and human-robot interaction communities, marking her as a rising voice in making reinforcement learning more accessible and responsive to human input.
Research Focus
Key Achievements
Top Papers
- 1Interactive Reinforcement Learning With Bayesian Fusion of Multimodal Advice12 citations · 2022