We examine a network of learners which address the same classification task but must learn from different data sets. The learners can share a limited portion of their data sets so as to preserve the network load. We introduce DELCO (standing for Decentralized Ensemble Learning with COpulas), a new approach in which the shared data and the trained models are sent to a central machine that allows to build an ensemble of classifiers. The proposed method aggregates the base classifiers using a probabilistic model relying on Gaussian copulas. Experiments on logistic regressor ensembles demonstrate competing accuracy and increased robustness as compared to gold standard approaches. A companion python implementation can be downloaded at https://github.com/john-klein/DELCO

We consider the stochastic shortest path (SSP) problem for succinct Markov decision processes (MDPs), where the MDP consists of a set of variables, and a set of nondeterministic rules that update the variables. First, we show that several examples from the AI literature can be modeled as succinct MDPs. Then we present computational approaches for upper and lower bounds for the SSP problem: (a) for computing upper bounds, our method is polynomial-time in the implicit description of the MDP; (b) for lower bounds, we present a polynomial-time (in the size of the implicit description) reduction to quadratic programming. Our approach is applicable even to infinite-state MDPs. Finally, we present experimental results to demonstrate the effectiveness of our approach on several classical examples from the AI literature.

A popular recent approach to answering open-domain questions is to first search for question-related passages and then apply reading comprehension models to extract answers. Existing methods usually extract answers from single passages independently. But some questions require a combination of evidence from across different sources to answer correctly. In this paper, we propose two models which make use of multiple passages to generate their answers. Both use an answer-reranking approach which reorders the answer candidates generated by an existing state-of-the-art QA model. We propose two methods, namely, strength-based re-ranking and coverage-based re-ranking, to make use of the aggregated evidence from different passages to better determine the answer. Our models have achieved state-of-the-art results on three public open-domain QA datasets: Quasar-T, SearchQA and the open-domain version of TriviaQA, with about 8 percentage points of improvement over the former two datasets.

I recently had a very interesting conversation with an air passenger whose flight had been delayed 12 hours by the "Beast from the East," a cold snap that was sweeping across Europe at the time, bringing some countries to a frozen standstill. During his long delay, the disgruntled flyer had heard nothing from his airline, nor had he been able to reach them through any of their communication channels. He even tried tweeting them. Knowing what I do for a living, he ended his tale with the question, "If they had artificial intelligence, would this fix their customer service?" The question illustrated how people often see artificial intelligence as a magical answer to all woes.

The European Emergency Number Association is expanding tests of an AI system that can help spot a heart attack. An AI tool that helps emergency call dispatchers detect a heart attack situation is set for wider testing in Europe. The Corti digital assistant, made by a company of the same name based in Copenhagen, Denmark, listens in on emergency calls and picks up on cues like breathing patterns, tone of voice and background noises to provide dispatchers with recommendations in real time. Corti is accurate in up to 95 percent of cases, according to the European Emergency Number Association, a Brussels-based nongovernmental organization devoted to improving Europe's emergency services, like its 112 distress call system. EENA said on Wednesday that it's teamed with Corti's creator to bring the system to four new sites throughout Europe for testing.

Two weeks ago I logged into good old Facebook dot com to discover I was one of the 311,127 Australians – and one of about 87 million people worldwide – who had their personal data harvested by Cambridge Analytica sometime around 2013-15. I was a small and unwitting cog in a vast, beguiling narrative of unfurling geopolitical upheaval encompassing the Trump presidency, Russian interference and Brexit. Here's what Facebook told me. I was not one of the 270,000-odd people who signed up to the now infamous This is Your Digital Life survey app but one of my friends was. As a result, Facebook "probably" shared my public profile, page likes, my date of birth and the city I lived in.

Pepper the robot certainly gets around. Besides welcoming folks at department stores and airports, the friendly android is also helping out at Pizza Hut and even working as a Buddhist priest. It recently put in a spell at a grocery store in Scotland, too, though admittedly that didn't work out so well. The diminutive droid's latest gig is at the Smithsonian in Washington, D.C. The organization claims to be "the first museum, research, and education complex in the world to experiment with this new and innovative technology," and is using 25 of the robots across its various locations in D.C. Built by Japanese telecoms giant SoftBank in collaboration with French robotics firm Aldebaran SAS, Pepper can recognize faces and emotions, and respond through voice or by showing messages and information on its torso-based tablet.

Google continues to add regional cloud partners as it seeks to differentiate its public cloud offerings while distributing its machine learning building blocks. Atos, the French big data platform and server vendor, announced a partnership with Google Cloud this week addressing secure hybrid cloud, data analytics and machine learning along with "digital workplace" initiatives. The partnership makes Google an Atos "preferred" cloud partner, the companies said Tuesday (April 24). Atos (EPA: ATO) said it would establish three machine learning and AI labs in France, U.K. and the U.S. that will use Google's training expertise to develop new machine learning models and applications. "Together, we will enable fast and smooth adoption of AI for enterprises," said Thierry Breton, chairman and CEO of Atos, Bezons, France.

The European Commission launched today the 3rd Data Economy Package aiming to assess emerging data issues to ensure a fair Digital Single Market. Also, the Communication on Artificial Intelligence (AI) was released – the EU's first step to look at the ethical questions surrounding AI. BusinessEurope Director General Markus J. Beyrer commented: "Comprehensive assessment is a good approach to foster digital technologies. Not every technological advance needs to be followed by legislation, only real market failures should be legislated for. The existing liability framework is fit for purpose. Although completely autonomous systems could need adapted rules in the future – we are not there yet. We need a balanced approach for the access of data. Openness is essential for the development of the digital economy, at the same time it is crucial that investments are protected. In our Digital Single Market, we prefer having contractual solutions because a number of parties with various roles are involved in the provision of digital services."

Brussels has called for a €20bn (£14bn) cash injection for artificial intelligence research, while pouring cold water over controversial plans to give robots human rights. The European commission wants governments and private companies to boost research and innovation spending on AI, amid rising concern Europe is losing ground to the US and China, where most leading AI firms are based. Health, transport and agriculture are among the areas the commission would like researchers to prioritise. But the commission distanced itself from proposals to give the most advanced robots the legal status of personhood. "I don't think it will happen," Andrus Ansip, a commission vice-president in charge of digital single-market policy told journalists.