Papers
1
Total Citations
28
H-Index
1
About
Fredrik Lindsten is a leading figure in computational statistics and machine learning, renowned for his pioneering work in sequential Monte Carlo (SMC) methods and probabilistic inference. His major contributions center on developing scalable, theoretically grounded algorithms for state-space models and Bayesian nonparametrics, with a particular focus on particle Markov chain Monte Carlo (PMCMC) and variational inference. In his highly cited work, "Sequential Kernel Herding: Frank-Wolfe Optimization for Particle Filtering" (2015, 28 citations), Lindsten elegantly bridges optimization and filtering by recasting particle filtering as a Frank-Wolfe algorithm, achieving superior sample efficiency over traditional Monte Carlo methods. This innovation exemplifies his broader impact: his research has garnered thousands of citations, with foundational papers on PMCMC and Rao-Blackwellized particle filters shaping modern inference pipelines. Lindsten’s ability to fuse rigorous theory with practical algorithms has made him a key architect of modern probabilistic programming, earning him recognition as a leading voice in the field. His work continues to inspire students and researchers tackling complex, high-dimensional inference problems.
Research Focus
Key Achievements
Top Papers
- 1Sequential Kernel Herding: Frank-Wolfe Optimization for Particle Filtering28 citations · 2015