We give a quasipolynomial-time algorithm for learning stochastic decision trees that is optimally resilient to adversarial noise. Given an $\eta$-corrupted set of uniform random samples labeled by a size-$s$ stochastic decision tree, our algorithm runs in time $n^{O(\log(s/\varepsilon)/\varepsilon^2)}$ and returns a hypothesis with error within an additive $2\eta + \varepsilon$ of the Bayes optimal. An additive $2\eta$ is the information-theoretic minimum. Previously no non-trivial algorithm with a guarantee of $O(\eta) + \varepsilon$ was known, even for weaker noise models. Our algorithm is furthermore proper, returning a hypothesis that is itself a decision tree; previously no such algorithm was known even in the noiseless setting.

Digital marketing optimization amounts to developing mechanisms which optimize the conversion of people using digital media to purchase products or services. The data coming from web sessions can be augmented with CRM matchings, from geographic positioning (from mobile data, IP address or other), from historical data and whatnot. Every industry and company has its own private stash of data and channels, it's in general difficult to describe a unique works-for-all approach. Technically this means a mixture of big-data techniques (Spark, MLLib, Hadoop…), R and Python coding, ML algorithms, dataviz through HTML/JS…and all of this together in a reusable, repeatable and collaborative workflow. This sounds like a daunting challenge but the good news is that the Dataiku platform makes it all really easy.