Crowdsourcing has become a popular paradigm for labeling large datasets. However, it has given rise to the computational task of aggregating the crowdsourced labels provided by a collection of unreliable annotators. We approach this problem by transforming it into a standard inference problem in graphical models, and applying approximate variational methods, including belief propagation (BP) and mean field (MF). We show that our BP algorithm generalizes both majority voting and a recent algorithm by Karger et al. [1], while our MF method is closely related to a commonly used EM algorithm. In both cases, we find that the performance of the algorithms critically depends on the choice of a prior distribution on the workers' reliability; by choosing the prior properly, both BP and MF (and EM) perform surprisingly well on both simulated and real-world datasets, competitive with state-of-the-art algorithms based on more complicated modeling assumptions.

We consider the two-group classification problem and propose a kernel classifier based on the optimal scoring framework. Unlike previous approaches, we provide theoretical guarantees on the expected risk consistency of the method. We also allow for feature selection by imposing structured sparsity using weighted kernels. We propose fully-automated methods for selection of all tuning parameters, and in particular adapt kernel shrinkage ideas for ridge parameter selection. Numerical studies demonstrate the superior classification performance of the proposed approach compared to existing nonparametric classifiers.

"When mistakes happen, patients die. I remember thinking to myself'What is going on here?' When Simon Kos practiced critical care medicine in intensive care and anaesthesia early on in his career, digital technology was not well established. "I saw mistakes – avoidable mistakes – happen all around me," he states. Now Chief Medical Officer at Microsoft, Kos is at the forefront of a medical revolution. His ambition is infectious, and his passion is fuelled by his first-hand experiences of the power that technological transformation, such as artificial intelligence (AI), can bring to our lives. Transforming healthcare The term AI is thrown around a lot these days. For many, it conjures up images of a dystopian future where robots have infiltrated society and have replaced our jobs – a perception born out of silver screen fantasies. "People think about artificial intelligence and immediately jump to the science fiction perceptions of robots," Kos states. "Instead, when you distill it, it's about using data in autonomous, mathematical ways, to deliver insights.

In a Cinco de Mayo weekend battle between high-profile Mexican boxers, Julio Cesar Chavez Jr. seeks an upset over Canelo Alvarez. Alvarez, one of the top pound-for-pound boxers, enters Saturday's fight at T-Mobile Arena in Las Vegas as 1/6 favorite. The fight is expected to have heavy attention consider Alvarez's reputation and Chavez's famous father. The two boxers, fighting at 164 pounds, are arguably the most well-known active Mexican boxers and they are set to tangle on Mexico's most famous holiday. While Chavez has never fought on a Cinco de Mayo weekend, Alvarez defeated Amir Khan at about the same time last year and knocked out James Kirkland on May 9, 2015, and also held off Shane Mosley on Cinco de Mayo in 2012.

Preview: Gennady "GGG" Golovkin (36-0, 33 KOs) faces challenger Daniel Jacobs (32-1, 29 KOs) in a middleweight unification bout Saturday night at Madison Square Garden in New York. Coverage of the fight will be on HBO pay-per-view. On the line are the WBA (Super), WBC, and IBO middleweight titles with two boxers entering the fight on knockout streaks. Golovkin, a native of Kazakhstan, has knocked out 23 opponents in a row. Jacobs, a native of Brooklyn, has 12 knockout victories in a row.

Police describe the area they are searching for the body of Ben Needham, close to where it is feared he may have been crushed to death by a digger. Ben, from Sheffield, was 21 months old when he disappeared on the Greek island of Kos on 24 July 1991. With the aid of drone footage, Det Insp John Cousins, South Yorkshire Police, explains the process.

Crowdsourcing has become a popular paradigm for labeling large datasets. However, ithas given rise to the computational task of aggregating the crowdsourced labels provided by a collection of unreliable annotators. We approach this problem bytransforming it into a standard inference problem in graphical models, and applying approximate variational methods, including belief propagation (BP) and mean field (MF). We show that our BP algorithm generalizes both majority votingand a recent algorithm by Karger et al. [1], while our MF method is closely related to a commonly used EM algorithm. In both cases, we find that the performance of the algorithms critically depends on the choice of a prior distribution onthe workers' reliability; by choosing the prior properly, both BP and MF (and EM) perform surprisingly well on both simulated and real-world datasets, competitive with state-of-the-art algorithms based on more complicated modeling assumptions.