Iterative Amortized Inference

Abstract

Inference models are a key component in scaling variational inference to deep latent variable models, most notably as encoder networks in variational auto-encoders (VAEs). By replacing conventional optimization-based inference with a learned model, inference is amortized over data examples and therefore more computationally efficient. However, standard inference models are restricted to direct mappings from data to approximate posterior estimates. The failure of these models to reach fully optimized approximate posterior estimates results in an amortization gap. We aim toward closing this gap by proposing iterative inference models, which learn to perform inference optimization through repeatedly encoding gradients. Our approach generalizes standard inference models in VAEs and provides insight into several empirical findings, including top-down inference techniques. We demonstrate the inference optimization capabilities of iterative inference models and show that they outperform standard inference models on several benchmark data sets of images and text.

arXiv commentInternational Conference on Machine Learning (ICML) 2018

Cited by in corpus (20)

• Recent advances and clinical applications of deep learning in medical image analysis
• Multi-Object Representation Learning with Iterative Variational Inference
• Semi-Amortized Variational Autoencoders
• Entity Abstraction in Visual Model-Based Reinforcement Learning
• A Tutorial on Deep Latent Variable Models of Natural Language
• NeRF-VAE: A Geometry Aware 3D Scene Generative Model
• Meta-Learning Deep Energy-Based Memory Models
• Preconditioned training of normalizing flows for variational inference in inverse problems
• Variational Laplace Autoencoders
• Regularization-Agnostic Compressed Sensing MRI Reconstruction with Hypernetworks
• Unsupervised Video Decomposition using Spatio-temporal Iterative Inference
• Better Set Representations For Relational Reasoning
• Control as Hybrid Inference
• Neural Function Modules with Sparse Arguments: A Dynamic Approach to Integrating Information across Layers
• Iterative VAE as a predictive brain model for out-of-distribution generalization
• Variational Auto-Decoder: A Method for Neural Generative Modeling from Incomplete Data
• Reinforcement Learning as Iterative and Amortised Inference
• Policy Message Passing: A New Algorithm for Probabilistic Graph Inference
• Unsupervised Representation Learning via Neural Activation Coding
• Accelerating Metropolis-Hastings with Lightweight Inference Compilation