On Combining Bags to Better Learn from Label Proportions

Rishi Saket, Aravindan Raghuveer, Balaraman Ravindran
Proceedings of The 25th International Conference on Artificial Intelligence and Statistics, PMLR 151:5913-5927, 2022.

Abstract

In the framework of learning from label proportions (LLP) the goal is to learn a good instance-level label predictor from the observed label proportions of bags of instances. Most of the LLP algorithms either explicitly or implicitly assume the nature of bag distributions with respect to the actual labels and instances, or cleverly adapt supervised learning techniques to suit LLP. In practical applications however, the scale and nature of data could render such assumptions invalid and the many of the algorithms impractical. In this paper we address the hard problem of solving LLP with provable error bounds while being bag distribution agnostic and model agnostic. We first propose the concept of generalized bags, an extension of bags and then devise an algorithm to combine bag distributions, if possible, into good generalized bag distributions. We show that (w.h.p) any classifier optimizing the squared Euclidean label-proportion loss on such a generalized bag distribution is guaranteed to minimize the instance-level loss as well. The predictive quality of our method is experimentally evaluated and it equals or betters the previous methods on pseudo-synthetic and real-world datasets.

Cite this Paper


BibTeX
@InProceedings{pmlr-v151-saket22a, title = { On Combining Bags to Better Learn from Label Proportions }, author = {Saket, Rishi and Raghuveer, Aravindan and Ravindran, Balaraman}, booktitle = {Proceedings of The 25th International Conference on Artificial Intelligence and Statistics}, pages = {5913--5927}, year = {2022}, editor = {Camps-Valls, Gustau and Ruiz, Francisco J. R. and Valera, Isabel}, volume = {151}, series = {Proceedings of Machine Learning Research}, month = {28--30 Mar}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v151/saket22a/saket22a.pdf}, url = {https://proceedings.mlr.press/v151/saket22a.html}, abstract = { In the framework of learning from label proportions (LLP) the goal is to learn a good instance-level label predictor from the observed label proportions of bags of instances. Most of the LLP algorithms either explicitly or implicitly assume the nature of bag distributions with respect to the actual labels and instances, or cleverly adapt supervised learning techniques to suit LLP. In practical applications however, the scale and nature of data could render such assumptions invalid and the many of the algorithms impractical. In this paper we address the hard problem of solving LLP with provable error bounds while being bag distribution agnostic and model agnostic. We first propose the concept of generalized bags, an extension of bags and then devise an algorithm to combine bag distributions, if possible, into good generalized bag distributions. We show that (w.h.p) any classifier optimizing the squared Euclidean label-proportion loss on such a generalized bag distribution is guaranteed to minimize the instance-level loss as well. The predictive quality of our method is experimentally evaluated and it equals or betters the previous methods on pseudo-synthetic and real-world datasets. } }
Endnote
%0 Conference Paper %T On Combining Bags to Better Learn from Label Proportions %A Rishi Saket %A Aravindan Raghuveer %A Balaraman Ravindran %B Proceedings of The 25th International Conference on Artificial Intelligence and Statistics %C Proceedings of Machine Learning Research %D 2022 %E Gustau Camps-Valls %E Francisco J. R. Ruiz %E Isabel Valera %F pmlr-v151-saket22a %I PMLR %P 5913--5927 %U https://proceedings.mlr.press/v151/saket22a.html %V 151 %X In the framework of learning from label proportions (LLP) the goal is to learn a good instance-level label predictor from the observed label proportions of bags of instances. Most of the LLP algorithms either explicitly or implicitly assume the nature of bag distributions with respect to the actual labels and instances, or cleverly adapt supervised learning techniques to suit LLP. In practical applications however, the scale and nature of data could render such assumptions invalid and the many of the algorithms impractical. In this paper we address the hard problem of solving LLP with provable error bounds while being bag distribution agnostic and model agnostic. We first propose the concept of generalized bags, an extension of bags and then devise an algorithm to combine bag distributions, if possible, into good generalized bag distributions. We show that (w.h.p) any classifier optimizing the squared Euclidean label-proportion loss on such a generalized bag distribution is guaranteed to minimize the instance-level loss as well. The predictive quality of our method is experimentally evaluated and it equals or betters the previous methods on pseudo-synthetic and real-world datasets.
APA
Saket, R., Raghuveer, A. & Ravindran, B.. (2022). On Combining Bags to Better Learn from Label Proportions . Proceedings of The 25th International Conference on Artificial Intelligence and Statistics, in Proceedings of Machine Learning Research 151:5913-5927 Available from https://proceedings.mlr.press/v151/saket22a.html.

Related Material