Deep Variational Bayes Filters: Unsupervised Learning of State Space Models from Raw Data

Abstract

We introduce Deep Variational Bayes Filters (DVBF), a new method for unsupervised learning and identification of latent Markovian state space models. Leveraging recent advances in Stochastic Gradient Variational Bayes, DVBF can overcome intractable inference distributions via variational inference. Thus, it can handle highly nonlinear input data with temporal and spatial dependencies such as image sequences without domain knowledge. Our experiments show that enabling backpropagation through transitions enforces state space assumptions and significantly improves information content of the latent embedding. This also enables realistic long-term prediction.

Published as a conference paper at ICLR 2017

References in corpus (3)

• Weight Uncertainty in Neural Networks
• Composing graphical models with neural networks for structured representations and fast inference
• Learning Stochastic Recurrent Networks

Cited by in corpus (32)

• KalmanNet: Neural Network Aided Kalman Filtering for Partially Known Dynamics
• State Representation Learning for Control: An Overview
• Dream to Control: Learning Behaviors by Latent Imagination
• Time Series Anomaly Detection for Cyber-Physical Systems via Neural System Identification and Bayesian Filtering
• ODE${}^2$VAE: Deep generative second order ODEs with Bayesian neural networks
• Recovering missing CFD data for high-order discretizations using deep neural networks and dynamics learning
• Physics-guided Deep Markov Models for Learning Nonlinear Dynamical Systems with Uncertainty
• Dissipative SymODEN: Encoding Hamiltonian Dynamics with Dissipation and Control into Deep Learning
• From Machine Learning to Robotics: Challenges and Opportunities for Embodied Intelligence
• Z-Forcing: Training Stochastic Recurrent Networks
• IONet: Learning to Cure the Curse of Drift in Inertial Odometry
• Interpretable End-to-end Urban Autonomous Driving with Latent Deep Reinforcement Learning
• Action and Perception as Divergence Minimization
• Recurrent Attentive Neural Process for Sequential Data
• Learning Dynamics Model in Reinforcement Learning by Incorporating the Long Term Future
• Disentangled State Space Representations
• Hierarchically Organized Latent Modules for Exploratory Search in Morphogenetic Systems
• Decentralized policy learning with partial observation and mechanical constraints for multiperson modeling
• Forecasting Individualized Disease Trajectories using Interpretable Deep Learning
• Automatic structured variational inference
• Vid2Param: Modelling of Dynamics Parameters from Video
• Improving Sequential Latent Variable Models with Autoregressive Flows
• Action-Sufficient State Representation Learning for Control with Structural Constraints
• On learning latent dynamics of the AUG plasma state
• Automatic variational inference with cascading flows
• A Worrying Analysis of Probabilistic Time-series Models for Sales Forecasting
• Stochastic Sequential Neural Networks with Structured Inference
• Deep Markov Spatio-Temporal Factorization
• A Generative Map for Image-based Camera Localization
• Learning Dynamical Systems from Noisy Sensor Measurements using Multiple Shooting
• Deformable Linear Object Prediction Using Locally Linear Latent Dynamics
• Switching Recurrent Kalman Networks