Calibrated Probabilistic Forecasts for Arbitrary Sequences

TL;DR

This paper introduces a forecasting framework that guarantees calibrated uncertainty estimates for any data sequence, regardless of distribution shifts or adversarial influences, using game theory principles.

Contribution

It presents a novel calibration framework based on Blackwell approachability, applicable to various prediction tasks, and extends it to recalibrate existing forecasters without losing accuracy.

Findings

Improved calibration in energy system forecasts

Effective recalibration of existing models

Robustness to distribution shifts and adversarial data

Abstract

Real-world data streams can change unpredictably due to distribution shifts, feedback loops and adversarial actors, which challenges the validity of forecasts. We present a forecasting framework ensuring valid uncertainty estimates regardless of how data evolves. Leveraging the concept of Blackwell approachability from game theory, we introduce a forecasting framework that guarantees calibrated uncertainties for outcomes in any compact space (e.g., classification or bounded regression). We extend this framework to recalibrate existing forecasters, guaranteeing calibration without sacrificing predictive performance. We implement both general-purpose gradient-based algorithms and algorithms optimized for popular special cases of our framework. Empirically, our algorithms improve calibration and downstream decision-making for energy systems.