Сравнение методов

	Динамический вариационный вывод	Фильтр Калмана
Область	Байесовские методы	Байесовские методы
Семейство	Bayesian methods	Bayesian methods
Год появления	2014–2015	1960
Автор метода	Bayer, Osendorfer, Krishnan and colleagues	Rudolf E. Kalman
Тип	Bayesian approximate inference	recursive Bayesian filter
Основополагающий источник	Krishnan, R. G., Shalit, U., & Sontag, D. (2015). Deep Kalman Filters. NIPS 2015 Workshop on Advances in Approximate Bayesian Inference.	Kalman, R. E. (1960). A new approach to linear filtering and prediction problems. Journal of Basic Engineering, 82(1), 35-45.
Другие названия	sequential variational inference, temporal variational inference, variational inference for state-space models, DVI	linear quadratic estimator, LQE, Kalman-Bucy filter, optimal recursive filter
Связанные	6	5
Сводка	Dynamic variational inference extends the variational inference framework to sequential and time-series settings by positing a structured approximate posterior that respects the temporal ordering of latent states. It jointly learns a generative model of how hidden states evolve over time and a recognition network that maps observed sequences back to those latent states, optimising a sequential evidence lower bound (ELBO).	The Kalman filter is an optimal recursive algorithm for estimating the hidden state of a linear dynamical system from noisy measurements. At each time step it alternates between a prediction step — projecting the state forward using the system model — and an update step that corrects the prediction with the new observation, producing minimum-variance state estimates and their uncertainty in real time.