Conghui Hu

About

Conghui Hu is a researcher specializing in probabilistic machine learning, state estimation, and sequential inference, with a particular focus on the intersection of deep learning and Bayesian filtering methods. Their most notable work centers on differentiable particle filters, where they have made significant contributions to bridging classical probabilistic inference with modern neural network architectures. In their influential 2021 paper, "End-to-End Semi-supervised Learning for Differentiable Particle Filters," Hu tackled a critical challenge in the field: enabling particle filters to scale to large-scale, real-world applications by making both dynamic and measurement models learnable through end-to-end training. This semi-supervised approach is particularly impactful, as it allows models to leverage both labeled and unlabeled data — a practical necessity in domains where annotation is expensive or scarce. With 15 citations, this work has already begun to shape how researchers think about combining the interpretability and theoretical guarantees of particle filters with the representational power of deep neural networks. Hu's research is highly relevant to robotics, autonomous systems, and any domain requiring robust sequential state estimation under uncertainty.

Research Focus

Key Achievements

Top Papers

1. 1
   End-to-End Semi-supervised Learning for Differentiable Particle Filters

   15 citations · 2021

Key Collaborators

Contact & Links