Based on a recent conversation between Joseph Sirosh, CTO for AI at Microsoft, and Roger Magoulas, VP of Radar at O’Reilly Media. Link to video recording below.

QMIX In many real-world settings, a team of agents must coordinate their behaviour while acting in a decentralised way. At the same time, it is often possible to train the agents in a centralised fashion in a simulated or laboratory setting, where global state information is available and communication constraints are lifted. Learning joint action-values conditioned on extra state information is an attractive way to exploit centralised learning, but the best strategy for then extracting decentralised policies is unclear. Our solution is QMIX, a novel value-based method that can train decentralised policies in a centralised end-to-end fashion. QMIX employs a network that estimates joint action-values as a complex non-linear combination of per-agent values that condition only on local observations. We structurally enforce that the joint-action value is monotonic in the per-agent values, which allows tractable maximisation of the joint action-value in off-policy learning, and guarantees consistency between the centralised and decentralised policies. We evaluate QMIX on a challenging set of StarCraft II micromanagement tasks, and show that QMIX significantly outperforms existing value-based multi-agent reinforcement learning methods. …

In case you haven’t heard, a whopping 85% of big data projects fail (we’ve already talked about why). But data science is an inherently risky endeavor, so the challenge today is not how to avoid failure completely, but how to fail in an efficient and productive way.

Softening Splits in Decision Trees (SplitSoftening)Allows to produce and use classification trees with probability (soft) splits.

Multitask Learning on Graph Neural Networks – Learning Multiple Graph Centrality Measures with a Unified Network

We recently saw a great recurring R question: “how do you use one column to choose a different value for each row?” That is: how do you use a column as an index? Please read on for some idiomatic base R, data.table, and dplyr solutions.

Hey, no one told me that baby name analysis was back in fashion. Dan Kopf for Quartz, using data from the Social Security Administration, describes the downfall of the name Heather. It exhibited the sharpest decline of all names since 1880.

Earlier this week, MIT, in collaboration with Boston Consulting Group, released their second global study looking at AI adoption in industry. A top finding of this report is that the leading companies in AI adoption are now convinced of the value of AI and are now facing the challenge of moving beyond individual point solutions toward broad, systematic use of AI across the company and at-scale.

Mini-batch Tempered MCMC (MINT-MCMC) In this paper we propose a general framework of performing MCMC with only a mini-batch of data. We show by estimating the Metropolis-Hasting ratio with only a mini-batch of data, one is essentially sampling from the true posterior raised to a known temperature. We show by experiments that our method, Mini-batch Tempered MCMC (MINT-MCMC), can efficiently explore multiple modes of a posterior distribution. As an application, we demonstrate one application of MINT-MCMC as an inference tool for Bayesian neural networks. We also show an cyclic version of our algorithm can be applied to build an ensemble of neural networks with little additional training cost. …

When Pol Brigneti started his career after studying Business Management at Pompeu Fabra University in Barcelona, he had no idea what data analytics was. After working as an intern for Ridelink he quickly discovered the importance of data analysis and realized that to acheive his goal of one day being a Product Manager he would need excellent data analytics skills.