Wesuggestanoveldistributional framework, able to represent arbitrary distribution functions over the continuous actionspace.

Uncertainty quantification is an important part of many performance critical applications. This paper provides a simple alternative to existing approaches such as ensemble learning and bayesian neural networks. By directly modeling the loss distribution with an Implicit Quantile Network, we get an estimate of how uncertain the model is of its predictions. For experiments with MNIST and CIFAR datasets, the mean of the estimated loss distribution is 2x higher for incorrect predictions. When data with high estimated uncertainty is removed from the test dataset, the accuracy of the model goes up as much as 10%. This method is simple to implement while offering important information to applications where the user has to know when the model could be wrong (e.g. deep learning for healthcare).

Importantly, our approach does not make Here, we propose a general method for probabilistic any a-priori assumptions on the underlying distribution of time series forecasting. We combine an our data. The probabilistic output of our model is generated autoregressive recurrent neural network to model via Implicit Quantile Networks (Dabney et al., 2018) temporal dynamics with Implicit Quantile Networks (IQN) and is trained by minimizing the integrand of the to learn a large class of distributions over a Continuous Ranked Probability Score (CRPS) (Matheson & time-series target. When compared to other probabilistic Winkler, 1976).

In this work, we build on recent advances in distributional reinforcement learning to give a generally applicable, flexible, and state-of-the-art distributional variant of DQN. We achieve this by using quantile regression to approximate the full quantile function for the state-action return distribution. By reparameterizing a distribution over the sample space, this yields an implicitly defined return distribution and gives rise to a large class of risk-sensitive policies. We demonstrate improved performance on the 57 Atari 2600 games in the ALE, and use our algorithm's implicitly defined distributions to study the effects of risk-sensitive policies in Atari games.