analiza aiqbench
This research addresses scalability and robustness in deep ensemble learning by introducing compact set-valued methods that balance storage efficiency with cautious, multi-utility predictions. By leveraging Bayesian Neural Networks and Monte-Carlo dropout to define representative distributions, the framework optimizes statistical distance for Bayes-optimal predictions while mitigating the risks of singleton outputs. The study provides empirical insights into the trade-offs between computational burden and prediction reliability in complex classification tasks.
Compact set-valued deep ensembling in multi-class classification
arXiv:2610.05332v1 Announce Type: new Abstract: This paper tackles visible challenges in deep ensemble learning, where deep neural networks serve as ensemble members: training and storage burdens, and robustness of cautious (set-valued) predictions targeting multiple utilities, which may involve reward-sensitivity. To mitigate the training and storage burdens, we propose to employ compact ensembles, such as Bayesian Neural Networks and Convolutional Neural Networks with the Monte-Carlo dropout prediction option, to produce probabilistic predictions. For each query instance, these probabilistic predictions are then used to define a representative distribution optimizing some statistical distance. The representative distribution is then employed to define the Bayes-optimal prediction (BOP) of any utility. To address the potential unrobustness of singleton prediction making, we propose a family of set-utilities satisfying some desirable properties and whose set-valued BOPs can be found efficiently. Empirical evidence is then given to illustrate the potential (dis)advantages of the proposed ensemble learning framework.
Abstract:This paper tackles visible challenges in deep ensemble learning, where deep neural networks serve as ensemble members: training and storage burdens, and robustness of cautious (set-valued) predictions targeting multiple utilities, which may involve reward-sensitivity. To mitigate the training and storage burdens, we propose to employ compact ensembles, such as Bayesian Neural Networks and Convolutional Neural Networks with the Monte-Carlo dropout prediction option, to produce probabilistic predictions. For each query instance, these probabilistic predictions are then used to define a representative distribution optimizing some statistical distance. The representative distribution is then employed to define the Bayes-optimal prediction (BOP) of any utility. To address the potential unrobustness of singleton prediction making, we propose a family of set-utilities satisfying some desirable properties and whose set-valued BOPs can be found efficiently. Empirical evidence is then given to illustrate the potential (dis)advantages of the proposed ensemble learning framework.
Machine Learning (cs.LG)
arXiv:2610.05332 [cs.LG]
(or arXiv:2610.05332v1 [cs.LG] for this version)
https://doi.org/10.48550/arXiv.2610.05332 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Text integral sub licența deschisă a publicației · 237 cuvinte
Subiecte conexe
[AINews] The Future of Latent Space
Latent Space · 25 sept. 2026
[AINews] Opus 5.5 is good at explainer videos
Latent Space · 29 sept. 2026
HyDI: A hybrid Deep Learning-Inductive Logic Programming ensemble for multi-label classification
arXiv · 30 sept. 2026
Handling Imbalanced Classification: What Works Better Than SMOTE
Analytics Vidhya · 12 iul. 2026