Adaptive Maximization of Pointwise Submodular Functions With Budget Constraint

Authors: Nguyen Cuong, Huan Xu

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

Reproducibility Variable Result LLM Response
Research Type Experimental We present experimental results for budgeted pool-based active learning with various modular cost settings. We use 3 binary classification data sets extracted from the 20 Newsgroups data [15]: alt.atheism/comp.graphics (data set 1), comp.sys.mac.hardware/comp.windows.x (data set 2), and rec.motorcycles/rec.sport.baseball (data set 3).
Researcher Affiliation Academia 1Department of Engineering, University of Cambridge, vcn22@cam.ac.uk 2Stewart School of Industrial & Systems Engineering, Georgia Institute of Technology, huan.xu@isye.gatech.edu
Pseudocode Yes Figure 1: Two greedy policies for adaptive optimization with budget constraint. Figure 2: The combined policy π1/2.
Open Source Code No The paper does not contain any statement or link indicating the public release of source code for the described methodology.
Open Datasets Yes We use 3 binary classification data sets extracted from the 20 Newsgroups data [15]: alt.atheism/comp.graphics (data set 1), comp.sys.mac.hardware/comp.windows.x (data set 2), and rec.motorcycles/rec.sport.baseball (data set 3).
Dataset Splits No The paper states using a 'separate test set' but does not provide specific details on train/validation/test splits, such as percentages, sample counts, or methods for creating these splits.
Hardware Specification No The paper does not provide any specific details about the hardware used for running experiments.
Software Dependencies No The paper mentions training a 'logistic regression model' but does not specify software dependencies with version numbers.
Experiment Setup Yes We train a logistic regression model with budgets 50, 100, 150, and 200, and approximate its area under the learning curve (AUC) using the accuracies on a separate test set. In this setting, costs are put randomly to the training examples in 2 scenarios. In scenario R1, some random examples have a cost drawn from Gamma(80, 0.1) and the other examples have cost 1. In scenario R2, all examples with label 1 have a cost drawn from Gamma(45, 0.1) and the others (examples with label 0) have cost 1. In scenario M1, we put higher costs on examples with lower margins. In scenario M2, we put higher costs on examples with larger margins.