Erik Derner
Papers
13
Total Citations
119
H-Index
7
About
Erik Derner’s research lies at the intersection of robotics, machine learning, and symbolic regression, with a focus on creating interpretable, data-driven models for dynamic systems. His major contributions include pioneering methods to construct parsimonious analytic models via symbolic regression, enabling human-comprehensible representations of complex behaviors—work that has garnered 25 citations. Derner has also advanced reinforcement learning by integrating symbolic input-output models, reducing the need for costly real-world trials in robot control. His work on change detection using weighted features for image-based localization (16 citations) and object-based pose graphs for dynamic indoor environments (12 citations) addresses critical challenges in lifelong robotic autonomy. Notably, his 2023 paper on a novel neural network approach to symbolic regression pushes toward physically plausible models, bridging data-driven learning with real-world interpretability. With over 100 total citations across his publications, Derner’s research is shaping how robots learn, adapt, and maintain robust performance in changing environments—a key step toward truly autonomous systems.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3Change detection using weighted features for image-based localization16 citations · 2020
- 4Object-Based Pose Graph for Dynamic Indoor Environments12 citations · 2020
- 5Reinforcement Learning with Symbolic Input-Output Models11 citations · 2018
- 6Efficient Object Search Through Probability-Based Viewpoint Selection10 citations · 2020
- 7
- 8
- 9
- 10