Alfredo Reichlin

Papers

Total Citations

H-Index

About

Alfredo Reichlin is a roboticist whose research lies at the intersection of machine learning, control, and physical interaction, with a particular focus on enabling robots to operate safely and dexterously in the real world. His most-cited work, "Back to the Manifold: Recovering from Out-of-Distribution States" (2022, 12 citations), addresses a critical challenge in learning from offline datasets: the distributional shift between training data and states encountered during deployment. Reichlin’s contribution is a framework that allows learned policies to recover from unfamiliar states, a key step toward safe, sample-efficient robot learning without costly online exploration. In parallel, his work on "Elastic Context: Encoding Elasticity for Data-driven Models of Textiles" (2022–2023, 4 citations each) tackles the complex problem of robotic manipulation of deformable objects. By encoding the material properties and construction techniques of textiles, Reichlin enables data-driven models to predict fabric behavior during stretching and pulling—a capability essential for assistive dressing and household tasks. His research bridges theory and application, advancing the frontier of robots that can learn from data and interact with the physical world in nuanced, human-like ways.