Masked Autoregressive Flow for Density Estimation
This work addresses density estimation for machine learning applications, offering a novel method that enhances model flexibility and performance.
The paper tackles the problem of improving flexibility in autoregressive models for density estimation by modeling internal random numbers through a stack of models, resulting in Masked Autoregressive Flow, which achieves state-of-the-art performance in general-purpose tasks.
Autoregressive models are among the best performing neural density estimators. We describe an approach for increasing the flexibility of an autoregressive model, based on modelling the random numbers that the model uses internally when generating data. By constructing a stack of autoregressive models, each modelling the random numbers of the next model in the stack, we obtain a type of normalizing flow suitable for density estimation, which we call Masked Autoregressive Flow. This type of flow is closely related to Inverse Autoregressive Flow and is a generalization of Real NVP. Masked Autoregressive Flow achieves state-of-the-art performance in a range of general-purpose density estimation tasks.