ssggreg/active_learning

class file : class_vogn.py

   implements the different Active learning methods

launch file : class_results_m.py

   specify which method to use, where to save, number of experiments, and hyperparameters

optimizer_.py : inherted class from optimizer used for a way to compute fast gradient length

active_learning

VOGN code : https://arxiv.org/abs/1807.04489

from:

@ARTICLE{2018arXiv180704489E, author = {{Emtiyaz Khan}, Mohammad and {Nielsen}, Didrik}, title = "{Fast yet Simple Natural-Gradient Descent for Variational Inference in Complex Models}", journal = {arXiv e-prints}, keywords = {Statistics - Machine Learning, Computer Science - Information Theory, Computer Science - Machine Learning, Statistics - Computation}, year = "2018", month = "Jul", eid = {arXiv:1807.04489}, pages = {arXiv:1807.04489}, archivePrefix = {arXiv}, eprint = {1807.04489}, primaryClass = {stat.ML}, adsurl = {https://ui.adsabs.harvard.edu/abs/2018arXiv180704489E}, adsnote = {Provided by the SAO/NASA Astrophysics Data System}