A New Approach to Linear Filtering and Prediction Problems

2002

Citations in the corpus

Recurrent Neural Filters: Learning Independent Bayesian Filtering Steps for Time Series Prediction (2019, methodology): “For many ﬁlter models – such as the Kalman ﬁlter (Kalman, 1960) and non-linear variants like the unscented Kalman ﬁlter (Julier & Uhlmann, 1997) – deterministic functions are used at each step to adjust the sufﬁcient statistics of the belief state, guided by generative models of the data.”
Financial Time Series Prediction Using Deep Learning (2017, methodology): “The Kalman Filter [1] and Auto-Regressive (AR) Models are seminal statistical approaches, where the ARMA [2] and ARIMA [3] (Auto-Regressive Integrated Moving Average) models are further generalizations of the AR, derived by combining the AR models with the moving-average (MA) models.”
Statistically optimized inversion algorithm for enhanced retrieval of aerosol properties from spectral multi-angle polarimetric satellite observations (2010, methodology): “…Squaretype formulation that complementarily unites advantages of a variety of practical inversion approaches, such asPhillips-Tikhonov-Twomeyconstrained inversion (Phillips, 1962; Tikhonov, 1963; Twomey, 1963), Kalman filter (Kalman, 1960),Newton-Gaussand Levenberg-Marquardtiterations, etc.”
Neural Survival Recommender (2017, methodology): “This is usually accomplished by a latent variable model, such as a Hidden Markov Model [29] or a nonlinear Kalman Filter [17].”
A General Method for Amortizing Variational Filtering (2018, methodology): “For linear Gaussian models, this manifests as the Kalman ﬁlter [26], which uses prediction errors to perform exact inference.”
Uncertainty in perception and the Hierarchical Gaussian Filter (2014, methodology): “Furthermore, while the Kalman ﬁlter (Kalman, 1960) is optimal for data generated by linear dynamical systems, the HGF has the advantage that it can deal with nonlinear systems because it adapts its volatility estimate as the data come in.”
Temporal Regularized Matrix Factorization for High-dimensional Time Series Prediction (2016, methodology): “Finally the DLM [7] is a latent AR(1) model with a general W (1) , which can be estimated by EM algorithms.”
Discrete-Time Nonlinear Filtering Algorithms Using Gauss–Hermite Quadrature (2007, methodology): “For this special case, the celebrated Kalman filter yields the optimal solution in the minimum-mean-square-error (MMSE) sense, the maximum likelihood (ML) sense, and the maximum a posteriori (MAP) sense [3], [18], [32].”
Active Neural Localization (2018, methodology): “These include classical methods based on Kalman Filters (Kalman et al., 1960; Smith et al., 1990) geometry-based visual odometry methods (Nist´er et al., 2006) and most recently, learning-based visual odometry methods which learn to predict motion between consecutive frames using recurrent convoluti”
Settable Systems: An Extension of Pearl's Causal Model with Optimization, Equilibrium, and Learning (2009, unknown): “hidden states, such as the Kalman lter (Kalman, 1960) and recurrent neural networks (e.”
TrackMate: An open and extensible platform for single-particle tracking. (2017, methodology): “A second particle-algorithm relies on the Kalman ﬁlter [17] to tackle linear motion.”
Entropy-based sensor selection heuristic for target localization (2004, methodology): “The Gaussian noise model has been widely assumed for sensor observations in many localization and tracking algorithms, e.g. the Kalman ﬁlter [9].”
Stochastic Skew in Currency Options (2007, unknown): “For both jumps, we use the same parameters (λ,vJ) ∈ R+ and α ≤ 2 for parsimony. This specification has its origin in the CGMY model of Carr, Ge man, Madan, and Yor (2002). We hence label it as CG jump. The Lévy density of the CG specification f llows an exponentially dampened power law. Depending on”
3D Traffic Scene Understanding From Movable Platforms (2014, unknown): “As the raw 3D location estimates { ( m , S ) are noisy due to the low camera viewpoint, the uncertainties in the object detector and the ground plane estimation process, we temporally integrate detections within a tracklet t us-ing a Kalman smoother [40] assuming constant velocity.”