# Is there a stochastic/online version of the GLM-Tron algorithm?

The GLM-Tron algorithm appeared in Theorem $$1$$ in this paper, https://arxiv.org/pdf/1104.2018.pdf Is there a stochastic version of this? (...essentially something that will randomly sample a few points of the data to construct the average vector that gets added to $$w_t$$ to get the next update $$w_{t+1}$$...)

Close cousins of GLM-Tron are the "Isotron" which has appeared here (https://www.microsoft.com/en-us/research/wp-content/uploads/2016/11/2009-The_Isotron_Algorithm.pdf) and the "Sparsitron" which has appeared here (https://arxiv.org/pdf/1706.06274.pdf) in Theorem $$3.1$$ All of these are also deterministic algorithms.

Except that Theorem $$3.1$$ in the above paper says that, "Moreover, the algorithm (Sparsitron) can be run in an online manner.". This seems quite ambiguous to me without a pseudocode being given for it. Sparsitron has 2 different loops and its not clear to me if there is an unique way to make an online version of it and if so whether the same proof given here will still apply.