Algorithmic Collective Action in Machine Learning

Moritz Hardt, Eric Mazumdar, Celestine Mendler-Dünner, Tijana Zrnic
Proceedings of the 40th International Conference on Machine Learning, PMLR 202:12570-12586, 2023.

Abstract

We initiate a principled study of algorithmic collective action on digital platforms that deploy machine learning algorithms. We propose a simple theoretical model of a collective interacting with a firm’s learning algorithm. The collective pools the data of participating individuals and executes an algorithmic strategy by instructing participants how to modify their own data to achieve a collective goal. We investigate the consequences of this model in three fundamental learning-theoretic settings: nonparametric optimal learning, parametric risk minimization, and gradient-based optimization. In each setting, we come up with coordinated algorithmic strategies and characterize natural success criteria as a function of the collective’s size. Complementing our theory, we conduct systematic experiments on a skill classification task involving tens of thousands of resumes from a gig platform for freelancers. Through more than two thousand model training runs of a BERT-like language model, we see a striking correspondence emerge between our empirical observations and the predictions made by our theory. Taken together, our theory and experiments broadly support the conclusion that algorithmic collectives of exceedingly small fractional size can exert significant control over a platform’s learning algorithm.

Cite this Paper

BibTeX
@InProceedings{pmlr-v202-hardt23a, title = {Algorithmic Collective Action in Machine Learning}, author = {Hardt, Moritz and Mazumdar, Eric and Mendler-D\"{u}nner, Celestine and Zrnic, Tijana}, booktitle = {Proceedings of the 40th International Conference on Machine Learning}, pages = {12570--12586}, year = {2023}, editor = {Krause, Andreas and Brunskill, Emma and Cho, Kyunghyun and Engelhardt, Barbara and Sabato, Sivan and Scarlett, Jonathan}, volume = {202}, series = {Proceedings of Machine Learning Research}, month = {23--29 Jul}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v202/hardt23a/hardt23a.pdf}, url = {https://proceedings.mlr.press/v202/hardt23a.html}, abstract = {We initiate a principled study of algorithmic collective action on digital platforms that deploy machine learning algorithms. We propose a simple theoretical model of a collective interacting with a firm’s learning algorithm. The collective pools the data of participating individuals and executes an algorithmic strategy by instructing participants how to modify their own data to achieve a collective goal. We investigate the consequences of this model in three fundamental learning-theoretic settings: nonparametric optimal learning, parametric risk minimization, and gradient-based optimization. In each setting, we come up with coordinated algorithmic strategies and characterize natural success criteria as a function of the collective’s size. Complementing our theory, we conduct systematic experiments on a skill classification task involving tens of thousands of resumes from a gig platform for freelancers. Through more than two thousand model training runs of a BERT-like language model, we see a striking correspondence emerge between our empirical observations and the predictions made by our theory. Taken together, our theory and experiments broadly support the conclusion that algorithmic collectives of exceedingly small fractional size can exert significant control over a platform’s learning algorithm.} }
Endnote
%0 Conference Paper %T Algorithmic Collective Action in Machine Learning %A Moritz Hardt %A Eric Mazumdar %A Celestine Mendler-Dünner %A Tijana Zrnic %B Proceedings of the 40th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2023 %E Andreas Krause %E Emma Brunskill %E Kyunghyun Cho %E Barbara Engelhardt %E Sivan Sabato %E Jonathan Scarlett %F pmlr-v202-hardt23a %I PMLR %P 12570--12586 %U https://proceedings.mlr.press/v202/hardt23a.html %V 202 %X We initiate a principled study of algorithmic collective action on digital platforms that deploy machine learning algorithms. We propose a simple theoretical model of a collective interacting with a firm’s learning algorithm. The collective pools the data of participating individuals and executes an algorithmic strategy by instructing participants how to modify their own data to achieve a collective goal. We investigate the consequences of this model in three fundamental learning-theoretic settings: nonparametric optimal learning, parametric risk minimization, and gradient-based optimization. In each setting, we come up with coordinated algorithmic strategies and characterize natural success criteria as a function of the collective’s size. Complementing our theory, we conduct systematic experiments on a skill classification task involving tens of thousands of resumes from a gig platform for freelancers. Through more than two thousand model training runs of a BERT-like language model, we see a striking correspondence emerge between our empirical observations and the predictions made by our theory. Taken together, our theory and experiments broadly support the conclusion that algorithmic collectives of exceedingly small fractional size can exert significant control over a platform’s learning algorithm.
APA
Hardt, M., Mazumdar, E., Mendler-Dünner, C. & Zrnic, T.. (2023). Algorithmic Collective Action in Machine Learning. Proceedings of the 40th International Conference on Machine Learning, in Proceedings of Machine Learning Research 202:12570-12586 Available from https://proceedings.mlr.press/v202/hardt23a.html.

Related Material