Credal Machine Learning for Risk-Averse Decision Making
Abstract
In many machine learning applications, it is necessary to guard against worst-case scenarios and predictions that could result in substantial losses. In principle, this can be achieved by training risk-averse predictive models that minimize loss functions such as conditional value-at-risk (CVaR), rather than relying on models that perform well on average. In practice, however, the effectiveness of this approach to risk aversion is undermined by the learner's uncertainty regarding the true loss distribution and, consequently, the true CVaR. To achieve reliable risk-aversion, we propose a method in which this (epistemic) uncertainty is represented in terms of credal sets, i.e., sets of probability distributions. More specifically, we develop an efficient yet reliable learner that produces predictions in the form of credal sets and combine it with a novel decision rule that maps each credal set to a single predictive distribution for CVaR minimization. Across classification, under distribution shift, and in reinforcement learning, our approach reliably avoids catastrophic decisions, while sacrificing little in expected performance.
Related
- Risk-Averse Decision Making with Multi-Level Reliability Guarantees
- Conformal Risk-Averse Decision Making with Optimized Certainty Equivalent Risk Control
- Risk-Averse Decision Making via Quantum Measurement Design
- Integrating Machine Learning and Uncertainty Quantification for Real-Time Decision-Making in Digital Twins: A System-Centric Review and Perspective
- Decision-Making in Digital Twins: Optimization, Control, and Learning
BibTeX
@misc{lohr2026credal,
title = {{Credal Machine Learning for Risk-Averse Decision Making}},
author = {Timo Löhr and Paul Hofman and Maximilian Muschalik and Eyke Hüllermeier},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2610.12115},
eprint = {2610.12115},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.12115},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - Credal Machine Learning for Risk-Averse Decision Making
AU - Timo Löhr
AU - Paul Hofman
AU - Maximilian Muschalik
AU - Eyke Hüllermeier
PY - 2026
DA - 2026/10/08
PB - arXiv
DO - 10.48550/arxiv.2610.12115
UR - https://arxiv.org/abs/2610.12115
AB - In many machine learning applications, it is necessary to guard against worst-case scenarios and predictions that could result in substantial losses. In principle, this can be achieved by training risk-averse predictive models that minimize loss functions such as conditional value-at-risk (CVaR), rather than relying on models that perform well on average. In practice, however, the effectiveness of this approach to risk aversion is undermined by the learner's uncertainty regarding the true loss distribution and, consequently, the true CVaR. To achieve reliable risk-aversion, we propose a method in which this (epistemic) uncertainty is represented in terms of credal sets, i.e., sets of probability distributions. More specifically, we develop an efficient yet reliable learner that produces predictions in the form of credal sets and combine it with a novel decision rule that maps each credal set to a single predictive distribution for CVaR minimization. Across classification, under distribution shift, and in reinforcement learning, our approach reliably avoids catastrophic decisions, while sacrificing little in expected performance.
ER -