Sam Toyer

Papers

Total Citations

H-Index

About

Sam Toyer’s research bridges computer vision and probabilistic machine learning, with a focus on human pose forecasting and deep generative models. His most-cited work, “Human Pose Forecasting via Deep Markov Models” (2017), introduces a novel framework that leverages deep Markov models to predict long-range 3D human skeleton sequences—a critical capability for applications in human-robot interaction, visual surveillance, and autonomous driving. By moving beyond short-term forecasts (milliseconds) to longer temporal horizons, Toyer’s approach addresses a key limitation in prior methods, enabling more robust and realistic motion prediction. This contribution has garnered 43 citations, reflecting its influence in advancing the field. Toyer’s work stands out for its integration of structured probabilistic reasoning with deep learning, offering a principled way to model uncertainty and temporal dependencies in human motion. His research not only pushes the boundaries of pose forecasting but also provides a foundation for safer, more responsive autonomous systems. For students and researchers, Toyer’s work exemplifies how combining classical statistical models with modern neural architectures can solve complex, real-world problems in embodied AI.