Svenja Stark

Papers

Total Citations

H-Index

About

Svenja Stark’s research lies at the intersection of robotics, motor skill learning, and autonomous adaptation, with a focus on enabling robots to learn and refine complex movements without human intervention. Her work addresses fundamental challenges in how machines acquire, store, and generalize motor skills—particularly in soft and musculoskeletal robotic systems. In her highly cited 2017 paper, Stark introduced a novel framework for autonomously learning nonparametric motor skill libraries, eliminating the need for expert-labeled data by comparing distance measures between new and stored experiences. She further advanced the field with her 2019 study on local online motor babbling, demonstrating how highly redundant musculoskeletal arms can learn motor abundance and solve inverse kinematics through goal-space exploration. Most recently, her 2020 work on model-based Quality-Diversity search pioneered a more efficient approach to generating diverse robot behaviors for open-ended manipulation tasks. Though early in her career, Stark’s cumulative work—garnering nearly a dozen citations across top venues—is shaping how robots autonomously build and leverage rich repertoires of motor skills, laying groundwork for more adaptive, lifelong learning in autonomous systems.