ContextFlow++: Generalist-Specialist Flow-based Generative Models with Mixed-Variable Context Encoding

TL;DR

ContextFlow++ introduces a novel flow-based generative model that effectively incorporates mixed-variable contexts with explicit generalist-specialist decoupling, leading to improved performance and training stability across diverse benchmarks.

Contribution

It proposes a new additive conditioning method with explicit decoupling and a mixed-variable architecture supporting discrete contexts, addressing limitations of traditional conditioning in flow models.

Findings

Faster stable training compared to previous models.

Higher performance metrics on multiple benchmarks.

Effective handling of discrete and continuous contexts.

Abstract

Normalizing flow-based generative models have been widely used in applications where the exact density estimation is of major importance. Recent research proposes numerous methods to improve their expressivity. However, conditioning on a context is largely overlooked area in the bijective flow research. Conventional conditioning with the vector concatenation is limited to only a few flow types. More importantly, this approach cannot support a practical setup where a set of context-conditioned (specialist) models are trained with the fixed pretrained general-knowledge (generalist) model. We propose ContextFlow++ approach to overcome these limitations using an additive conditioning with explicit generalist-specialist knowledge decoupling. Furthermore, we support discrete contexts by the proposed mixed-variable architecture with context encoders. Particularly, our context encoder for discrete variables is a surjective flow from which the context-conditioned continuous variables are sampled. Our experiments on rotated MNIST-R, corrupted CIFAR-10C, real-world ATM predictive maintenance and SMAP unsupervised anomaly detection benchmarks show that the proposed ContextFlow++ offers faster stable training and achieves higher performance metrics. Our code is publicly available at https://github.com/gudovskiy/contextflow.