This paper introduces a novel framework for deep ensemble learning in multi-class classification tasks. It addresses challenges related to training efficiency and storage by proposing the use of compact ensembles, such as Bayesian Neural Networks and Convolutional Neural Networks with Monte Carlo Dropout. These methods generate probabilistic predictions that are then used to define a representative distribution, optimizing statistical distance. This distribution helps find the Bayes-optimal prediction (BOP) for various utilities, including a proposed family of set-utilities designed for robust, set-valued BOPs. AI
IMPACT Introduces a new method for improving efficiency and robustness in deep ensemble learning for classification tasks.
RANK_REASON The cluster contains a research paper submitted to arXiv detailing a new methodology for deep ensemble learning. [lever_c_demoted from research: ic=1 ai=1.0]
- arXiv
- Bayesian Neural Networks
- convolutional neural network
- Hugging Face
- Kim-Dung Tran
- Monte Carlo Dropout
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →