mlsanityData to 7 Oct 2026

Unsupervised Learned Kalman Filtering

arXiv 2110.09005 ↗First posted 18 Oct 2021eess.SP cs.LGcited by 17

Not on the map. Why?

Abstract (arXiv)

In this paper we adapt KalmanNet, which is a recently pro-posed deep neural network (DNN)-aided system whose architecture follows the operation of the model-based Kalman filter (KF), to learn its mapping in an unsupervised manner, i.e., without requiring ground-truth states. The unsupervised adaptation is achieved by exploiting the hybrid model-based/data-driven architecture of KalmanNet, which internally predicts the next observation as the KF does. These internal features are then used to compute the loss rather than the state estimate at the output of the system. With the capability of unsupervised learning, one can use KalmanNet not only to track the hidden state, but also to adapt to variations in the state space (SS) model. We numerically demonstrate that when the noise statistics are unknown, unsupervised KalmanNet achieves a similar performance to KalmanNet with supervised learning. We also show that we can adapt a pre-trained KalmanNet to changing SS models without providing additional data thanks to the unsupervised capabilities.

Key facts. Cited by 17 papers in the map (data through 7 Oct 2026). Builds on 17 of its 28 references that are in the map. First posted on arXiv 18 Oct 2021.

Builds on

17 of 28 references.

References · 17

  1. 2016Structured Inference Networks for Nonlinear State Space Modelscited by 182
  2. 2015Deep Kalman Filterscited by 163
  3. 2016Deep Variational Bayes Filters: Unsupervised Learning of State Space Models from Raw Datacited by 154
  4. 2020Model-Based Deep Learningcited by 136
  5. 2021KalmanNet: Neural Network Aided Kalman Filtering for Partially Known Dynamicscited by 128
  6. 2017A Disentangled Recognition and Nonlinear Dynamics Model for Unsupervised Learningcited by 107
  7. 2017Variational Sequential Monte Carlocited by 93
  8. 2021Chaos as an interpretable benchmark for forecasting and data-driven modellingcited by 87
  9. 2019Recurrent Kalman Networks: Factorized Inference in High-Dimensional Deep Feature Spacescited by 64
  10. 2016Backprop KF: Learning Discriminative Deterministic State Estimatorscited by 47
  11. 2020KFNet: Learning Temporal Camera Relocalization using Kalman Filteringcited by 43
  12. 2017Long Short-Term Memory Kalman Filters:Recurrent Neural Estimators for Pose Regularizationcited by 36
5 more
  1. 2019Combining Generative and Discriminative Models for Hybrid Inferencecited by 34
  2. 2015Black box variational inference for state space modelscited by 31
  3. 2019ViterbiNet: A Deep Learning Based Viterbi Algorithm for Symbol Detectioncited by 25
  4. 2020DeepSIC: Deep Soft Interference Cancellation for Multiuser MIMO Detectioncited by 23
  5. 2017State Space LSTM Models with Particle MCMC Inferencecited by 11

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2022: 2; 2023: 4; 2024: 6; 2025: 2; 2026: 3.

17 papers.

Sort:

  1. 2023DANSE: Data-driven Non-linear State Estimation of Model-free Process in Unsupervised Learning Setupcited by 19
  2. 2023Radars for Autonomous Driving: A Review of Deep Learning Methods and Challengescited by 10
  3. 2023GSP-KalmanNet: Tracking Graph Signals via Neural-Aided Kalman Filteringcited by 10
  4. 2024Normalizing Flow-based Differentiable Particle Filterscited by 10
  5. 2023SubspaceNet: Deep Learning-Aided Subspace Methods for DoA Estimationcited by 7
  6. 2024A Survey of Machine Learning Techniques for Improving Global Navigation Satellite Systemscited by 4
  7. 2024AI-Aided Kalman Filterscited by 3
  8. 2022Neural Augmented Kalman Filtering with Bollinger Bands for Pairs Tradingcited by 2
  9. 2024Hybrid Data-Driven SSM for Interpretable and Label-Free mmWave Channel Predictioncited by 2
  10. 2024Semi-Supervised Model-Free Bayesian State Estimation from Compressed Measurementscited by 1
  11. 2022LQGNet: Hybrid Model-Based and Data-Driven Linear Quadratic Stochastic Controlcited by 0
  12. 2024Learning Flock: Enhancing Sets of Particles for Multi~Sub-State Particle Filtering with Neural Augmentationcited by 0
5 more
  1. 2025DIMM: Decoupled Multi-hierarchy Kalman Filter for 3D Object Trackingcited by 0
  2. 2025pDANSE: Particle-based Data-driven Nonlinear State Estimation from Nonlinear Measurementscited by 0
  3. 2026VSE: Variational state estimation of complex model-free processcited by 0
  4. 2026RTS Smoother-Guided Learning of Physics-Based Neural Differential Modelscited by 0
  5. 2026Unscented KalmanNet: Structure-Preserving Deep Learning with Calibrated Posterior Uncertainty under Incomplete Physics and Unknown Noisecited by 0