Learning the Truth From Only One Side of the Story

Heinrich Jiang, Qijia Jiang, Aldo Pacchiano
Proceedings of The 24th International Conference on Artificial Intelligence and Statistics, PMLR 130:2413-2421, 2021.

Abstract

Learning under one-sided feedback (i.e., where we only observe the labels for examples we predicted positively on) is a fundamental problem in machine learning – applications include lending and recommendation systems. Despite this, there has been surprisingly little progress made in ways to mitigate the effects of the sampling bias that arises. We focus on generalized linear models and show that without adjusting for this sampling bias, the model may converge suboptimally or even fail to converge to the optimal solution. We propose an adaptive approach that comes with theoretical guarantees and show that it outperforms several existing methods empirically. Our method leverages variance estimation techniques to efficiently learn under uncertainty, offering a more principled alternative compared to existing approaches.

Cite this Paper


BibTeX
@InProceedings{pmlr-v130-jiang21b, title = { Learning the Truth From Only One Side of the Story }, author = {Jiang, Heinrich and Jiang, Qijia and Pacchiano, Aldo}, booktitle = {Proceedings of The 24th International Conference on Artificial Intelligence and Statistics}, pages = {2413--2421}, year = {2021}, editor = {Banerjee, Arindam and Fukumizu, Kenji}, volume = {130}, series = {Proceedings of Machine Learning Research}, month = {13--15 Apr}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v130/jiang21b/jiang21b.pdf}, url = {https://proceedings.mlr.press/v130/jiang21b.html}, abstract = { Learning under one-sided feedback (i.e., where we only observe the labels for examples we predicted positively on) is a fundamental problem in machine learning – applications include lending and recommendation systems. Despite this, there has been surprisingly little progress made in ways to mitigate the effects of the sampling bias that arises. We focus on generalized linear models and show that without adjusting for this sampling bias, the model may converge suboptimally or even fail to converge to the optimal solution. We propose an adaptive approach that comes with theoretical guarantees and show that it outperforms several existing methods empirically. Our method leverages variance estimation techniques to efficiently learn under uncertainty, offering a more principled alternative compared to existing approaches. } }
Endnote
%0 Conference Paper %T Learning the Truth From Only One Side of the Story %A Heinrich Jiang %A Qijia Jiang %A Aldo Pacchiano %B Proceedings of The 24th International Conference on Artificial Intelligence and Statistics %C Proceedings of Machine Learning Research %D 2021 %E Arindam Banerjee %E Kenji Fukumizu %F pmlr-v130-jiang21b %I PMLR %P 2413--2421 %U https://proceedings.mlr.press/v130/jiang21b.html %V 130 %X Learning under one-sided feedback (i.e., where we only observe the labels for examples we predicted positively on) is a fundamental problem in machine learning – applications include lending and recommendation systems. Despite this, there has been surprisingly little progress made in ways to mitigate the effects of the sampling bias that arises. We focus on generalized linear models and show that without adjusting for this sampling bias, the model may converge suboptimally or even fail to converge to the optimal solution. We propose an adaptive approach that comes with theoretical guarantees and show that it outperforms several existing methods empirically. Our method leverages variance estimation techniques to efficiently learn under uncertainty, offering a more principled alternative compared to existing approaches.
APA
Jiang, H., Jiang, Q. & Pacchiano, A.. (2021). Learning the Truth From Only One Side of the Story . Proceedings of The 24th International Conference on Artificial Intelligence and Statistics, in Proceedings of Machine Learning Research 130:2413-2421 Available from https://proceedings.mlr.press/v130/jiang21b.html.

Related Material