Joint blind source separation (J-BSS) is an emerging data-driven technique for multi-set data-fusion. In this paper, J-BSS is addressed from a tensorial perspective. We show how, by using second-order multi-set statistics in J-BSS, a specific double coupled canonical polyadic decomposition (DC-CPD) problem can be formulated. We propose an algebraic DC-CPD algorithm based on a coupled rank-1 detection mapping. This algorithm converts a possibly underdetermined DC-CPD to a set of overdetermined CPDs. The latter can be solved algebraically via a generalized eigenvalue decomposition based scheme. Therefore, this algorithm is deterministic and returns the exact solution in the noiseless case. In the noisy case, it can be used to effectively initialize optimization based DC-CPD algorithms. In addition, we obtain the determini- stic and generic uniqueness conditions for DC-CPD, which are shown to be more relaxed than their CPD counterpart. Experiment results are given to illustrate the superiority of DC-CPD over standard CPD based BSS methods and several existing J-BSS methods, with regards to uniqueness and accuracy.

We introduce a new family of MCMC samplers that combine auxiliary variables, Gibbs sampling and Taylor expansions of the target density. Our approach permits the marginalisation over the auxiliary variables yielding marginal samplers, or the augmentation of the auxiliary variables, yielding auxiliary samplers. The well-known Metropolis-adjusted Langevin algorithm (MALA) and preconditioned Crank-Nicolson Langevin (pCNL) algorithm are shown to be special cases. We prove that marginal samplers are superior in terms of asymptotic variance and demonstrate cases where they are slower in computing time compared to auxiliary samplers. In the context of latent Gaussian models we propose new auxiliary and marginal samplers whose implementation requires a single tuning parameter, which can be found automatically during the transient phase. Extensive experimentation shows that the increase in efficiency (measured as effective sample size per unit of computing time) relative to (optimised implementations of) pCNL, elliptical slice sampling and MALA ranges from 10-fold in binary classification problems to 25-fold in log-Gaussian Cox processes to 100-fold in Gaussian process regression, and it is on par with Riemann manifold Hamiltonian Monte Carlo in an example where the latter has the same complexity as the aforementioned algorithms. We explain this remarkable improvement in terms of the way alternative samplers try to approximate the eigenvalues of the target. We introduce a novel MCMC sampling scheme for hyperparameter learning that builds upon the auxiliary samplers. The MATLAB code for reproducing the experiments in the article is publicly available and a Supplement to this article contains additional experiments and implementation details.

We propose Rademacher complexity bounds for multiclass classifiers trained with a two-step semi-supervised model. In the first step, the algorithm partitions the partially labeled data and then identifies dense clusters containing $\kappa$ predominant classes using the labeled training examples such that the proportion of their non-predominant classes is below a fixed threshold. In the second step, a classifier is trained by minimizing a margin empirical loss over the labeled training set and a penalization term measuring the disability of the learner to predict the $\kappa$ predominant classes of the identified clusters. The resulting data-dependent generalization error bound involves the margin distribution of the classifier, the stability of the clustering technique used in the first step and Rademacher complexity terms corresponding to partially labeled training data. Our theoretical result exhibit convergence rates extending those proposed in the literature for the binary case, and experimental results on different multiclass classification problems show empirical evidence that supports the theory.

"AI is probably the most important thing humanity has ever worked on. I think of it as something more profound than electricity or fire." That's a bold claim for anyone to make, but even more significant given the speaker: Google (GOOGL) boss Sundar Pichai. He was giving his take on the frontier technology in a speech at the World Economic Forum in Davos, Switzerland, on Wednesday. Pichai, Google CEO since 2015, has pushed his firm deeper into developing artificial intelligence under parent company Alphabet (GOOGL).

The prime minister is to say she wants the UK to lead the world in deciding how artificial intelligence can be deployed in a safe and ethical manner. Theresa May will say at the World Economic Forum in Davos that a new advisory body, previously announced in the Autumn Budget, will co-ordinate efforts with other countries. In addition, she will confirm that the UK will join the Davos forum's own council on artificial intelligence. But others may have stronger claims. Earlier this week, Google picked France as the base for a new research centre dedicated to exploring how AI can be applied to health and the environment.

How do we stop artificial intelligence from going rogue? The idea is scary enough that robotics companies are proposing that the UN put a ban on killer autonomous robots. And it's made even scarier by mounting evidence that engineers actually understand very little about how AI algorithms do what they do. Doomsday singularity scenarios aside, rogue AI presents a very serious problem even in more everyday terms. What if the autonomous cars chauffeuring us around reach the wrong conclusions on how they should operate in traffic?

Brightcrowd -- helps you find meaningful professional connections Capsule.ai Abi -- your virtual health assistant Ada -- can help if you're feeling unwell Airi -- personal health coach Alz.ai -- helps you care for loved ones with Alzheimer's Amélie -- chatbot for mental health Bitesnap -- food recognition from meal photos to help count calories doc.ai -- makes lab results easy to understand Gyan -- helps you go from symptoms to likely conditions Joy -- helps you track and improve your mental health Kiwi -- helps you to reduce and quit smoking Tess by X2AI-- therapist in your pocket Sleep.ai Amazon Echo / Alexa -- everyday personal assistant for in-home Apple Siri -- everyday personal assistant on iPhone and Mac Cortana-- everyday personal assistant on PC and Windows devices Facebook M -- competitors to Siri, Now and Cortana Focus -- helps you focus, get tasks done and prioritise your day Gatebox -- a holographic anime assistant in an espresso machine Google Assistant -- everyday personal assistant Hound -- everyday personal assistant Ling -- similar to Amazon Echo Mycroft -- is the world's first open source voice assistant Remi-- like Siri with an interface Spoken -- virtual assistant with an interface Viv -- like Siri but 10x better Clara -- meeting scheduling assistant Julie Desk -- meeting scheduling assistant (aimed at C-Suite) Kono -- meeting scheduling assistant Mimetic -- meeting scheduling assistant My Ally -- handles meeting scheduling and manages calendar SkipFlag -- automatically discover and organise your work Vesper -- virtual assistant aimed at C-Suite x.ai-- meeting scheduling assistant Zoom.ai -- personal assistant to help you at work Jottr -- content and news app that learns what you like and don't like News360 -- learns what you enjoy and finds stories you'll like Entrupy -- helps detect if high-end designer products are authentic Fify -- helps you shop for clothing GoFind -- helps you find clothing online by taking a photo Mode.ai -- helps you find clothing online Abe -- fast answers about your finances Andy -- a personal Tax Accountant Ara -- helps you budget Bond -- helps you achieve your financial goals Mylo -- rounds up your everyday purchases and invest the spare change Olivia -- helps you manage your finances Responsive-- institutional-grade active portfolio management Roger -- helps you pay bills easily Wallet -- AI for your daily finance decisions Xoe.ai -- AI lending chatbot Firedrop -- websites designed automatically, just add content and publish Hashley -- ironic hashtag and comment generator for your photos Millions.ai Aerial -- home activity, movement and identity sensor Bridge.ai Eli -- helps you learn a new language from conversations through your day Kick.ai

Wed 24 Jan 2018 14.02 EST Last modified on Wed 24 Jan 2018 14.15 EST The chief executive of Google has declared he is happy for his company to pay more tax, and called for the existing system to be reformed. Sundar Pichai told an audience at the World Economic Forum in Davos that the tax system needed to be reformed to address concerns that some companies were not paying their fair share. Speaking before the French president, Emmanuel Macron, challenged tech giants to pay more tax, Pichai said: "As a company we paid, over the last five years, close to 20% in tax. We are happy to pay a higher amount, whatever the world agrees on as the right framework. It's not an issue about the amount of tax we pay, as much as how you divide it among various countries."

Life is changing fast for manufacturing companies. Not only is the business of designing, making, shipping, distributing and selling finished goods becoming ever more global and competitive, but technology is ensuring that it is evolving at a pace that's tough to keep up with. Manufacturing companies need to be more agile, thoughtful and innovative than ever in how they do business. Manufacturing is a connected, but often global, distributed operation. Many of today's complex products use raw materials and components from all over the world in their manufacture.

Wearable technology that can predict cancerous growths'several months' before they form are just around the corner, according to one technology expert. Nokia's chief says the firm is working on a scanning device that will pick up on biomarkers that indicate the conditions needed for abnormal cell growth to happen. A number of other medical innovations have also been envisaged that will make use of ultra-fast 5G mobile internet networks in the future. That includes remote surgery conducted from across the world, as well as ambulances that are able to transmit data to a hospital ahead of its arrival. Wearable technology that can predict cancerous growths'several months' before they form are just around the corner, according Nokia's chief Rajeev Suri.