Controlled abstention neural networks for identifying skillful predictions for classification problems

Abstract

The earth system is exceedingly complex and often chaotic in nature, making prediction incredibly challenging: we cannot expect to make perfect predictions all of the time. Instead, we look for specific states of the system that lead to more predictable behavior than others, often termed "forecasts of opportunity." When these opportunities are not present, scientists need prediction systems that are capable of saying "I don't know." We introduce a novel loss function, termed the "NotWrong loss", that allows neural networks to identify forecasts of opportunity for classification problems. The NotWrong loss introduces an abstention class that allows the network to identify the more confident samples and abstain (say "I don't know") on the less confident samples. The abstention loss is designed to abstain on a user-defined fraction of the samples via a PID controller. Unlike many machine learning methods used to reject samples post-training, the NotWrong loss is applied during training to preferentially learn from the more confident samples. We show that the NotWrong loss outperforms other existing loss functions for multiple climate use cases. The implementation of the proposed loss function is straightforward in most network architectures designed for classification as it only requires the addition of an abstention class to the output layer and modification of the loss function.

arXiv commentsubmitted to the Journal of Advances in Earth System Modeling. arXiv admin note: substantial text overlap with arXiv:2104.08236

References in corpus (11)

• Explaining NonLinear Classification Decisions with Deep Taylor Decomposition
• Physically Interpretable Neural Networks for the Geosciences: Applications to Earth System Variability
• Machine Learning for Stochastic Parameterization: Generative Adversarial Networks in the Lorenz '96 Model
• Stochastic Super-Resolution for Downscaling Time-Evolving Atmospheric Fields with a Generative Adversarial Network
• Neural Network Attribution Methods for Problems in Geoscience: A Novel Synthetic Benchmark Dataset
• Combating Label Noise in Deep Learning Using Abstention
• SelectiveNet: A Deep Neural Network with an Integrated Reject Option
• Controlled abstention neural networks for identifying skillful predictions for regression problems
• Resampling with neural networks for stochastic parameterization in multiscale systems
• Identifying Opportunities for Skillful Weather Prediction with Interpretable Neural Networks
• A Unified Plug-and-Play Framework for Effective Data Denoising and Robust Abstention

Cited by in corpus (3)

• Distributional neural networks for electricity price forecasting
• Controlled abstention neural networks for identifying skillful predictions for regression problems
• CIRA Guide to Custom Loss Functions for Neural Networks in Environmental Sciences -- Version 1