Dropout Strikes Back: Improved Uncertainty Estimation via Diversity Sampling
This improves uncertainty estimation for applications like confidence intervals and out-of-distribution detection, offering a straightforward enhancement for deep learning models with dropout layers.
The paper tackles uncertainty estimation in neural networks by modifying dropout sampling distributions to include maximally diverse neurons, achieving state-of-the-art results in regression and classification tasks without model modifications.
Uncertainty estimation for machine learning models is of high importance in many scenarios such as constructing the confidence intervals for model predictions and detection of out-of-distribution or adversarially generated points. In this work, we show that modifying the sampling distributions for dropout layers in neural networks improves the quality of uncertainty estimation. Our main idea consists of two main steps: computing data-driven correlations between neurons and generating samples, which include maximally diverse neurons. In a series of experiments on simulated and real-world data, we demonstrate that the diversification via determinantal point processes-based sampling achieves state-of-the-art results in uncertainty estimation for regression and classification tasks. An important feature of our approach is that it does not require any modification to the models or training procedures, allowing straightforward application to any deep learning model with dropout layers.