Optimal Binary Classifier Aggregation for General Losses

Authors: Akshay Balsubramani, Yoav S. Freund

NeurIPS 2016 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Theoretical We address the problem of aggregating an ensemble of predictors with known loss bounds in a semi-supervised binary classification setting, to minimize prediction loss incurred on the unlabeled data. We find the minimax optimal predictions for a very general class of loss functions including all convex and many non-convex losses, extending a recent analysis of the problem for misclassification error. The result is a family of semi-supervised ensemble aggregation algorithms which are as efficient as linear learning by convex optimization, but are minimax optimal without any relaxations. Their decision rules take a form familiar in decision theory applying sigmoid functions to a notion of ensemble margin without the assumptions typically made in margin-based learning.
Researcher Affiliation Academia Akshay Balsubramani University of California, San Diego abalsubr@ucsd.edu Yoav Freund University of California, San Diego yfreund@ucsd.edu
Pseudocode No The paper describes an 'Ensemble Aggregation Algorithm' in Section 2.2, but it is presented as descriptive text rather than a formal pseudocode block or algorithm box.
Open Source Code No The paper does not provide any explicit statements about open-source code availability, nor does it include links to code repositories.
Open Datasets No The paper is theoretical and does not present experiments using specific datasets. It discusses 'training set' in a general problem context, but not as a dataset used for its own evaluation.
Dataset Splits No The paper is theoretical and does not conduct experiments; therefore, it does not discuss validation dataset splits.
Hardware Specification No The paper is theoretical and does not describe any hardware used for experiments.
Software Dependencies No The paper is theoretical and does not mention specific software dependencies with version numbers for implementation.
Experiment Setup No The paper is theoretical and does not describe any experimental setup details such as hyperparameters or training configurations.