Matrix Factorization is a popular non-convex objective, for which alternating minimization schemes are mostly used. They usually suffer from the major drawback that the solution is biased towards one of the optimization variables. A remedy is non-alternating schemes. However, due to a lack of Lipschitz continuity of the gradient in matrix factorization problems, convergence cannot be guaranteed. A recently developed remedy relies on the concept of Bregman distances, which generalizes the standard Euclidean distance. We exploit this theory by proposing a novel Bregman distance for matrix factorization problems, which, at the same time, allows for simple/closed form update steps. Therefore, for non-alternating schemes, such as the recently introduced Bregman Proximal Gradient (BPG) method and an inertial variant Convex--Concave Inertial BPG (CoCaIn BPG), convergence of the whole sequence to a stationary point is proved for Matrix Factorization. In several experiments, we observe a superior performance of our non-alternating schemes in terms of speed and objective value at the limit point.

We propose a scalable Gromov-Wasserstein learning (S-GWL) method and establish a novel and theoretically-supported paradigm for large-scale graph analysis. The proposed method is based on the fact that Gromov-Wasserstein discrepancy is a pseudometric on graphs. Given two graphs, the optimal transport associated with their Gromov-Wasserstein discrepancy provides the correspondence between their nodes and achieves graph matching. When one of the graphs has isolated but self-connected nodes ($i.e.$, a disconnected graph), the optimal transport indicates the clustering structure of the other graph and achieves graph partitioning. Using this concept, we extend our method to multi-graph partitioning and matching by learning a Gromov-Wasserstein barycenter graph for multiple observed graphs; the barycenter graph plays the role of the disconnected graph, and since it is learned, so is the clustering. Our method combines a recursive $K$-partition mechanism with a regularized proximal gradient algorithm, whose time complexity is $\mathcal{O}(K(E+V)\log_K V)$ for graphs with $V$ nodes and $E$ edges. To our knowledge, our method is the first attempt to make Gromov-Wasserstein discrepancy applicable to large-scale graph analysis and unify graph partitioning and matching into the same framework. It outperforms state-of-the-art graph partitioning and matching methods, achieving a trade-off between accuracy and efficiency.

This paper provides a unifying theoretical framework for stochastic optimization algorithms by means of a latent stochastic variational problem. Using techniques from stochastic control, the solution to the variational problem is shown to be equivalent to that of a Forward Backward Stochastic Differential Equation (FBSDE). By solving these equations, we recover a variety of existing adaptive stochastic gradient descent methods. This framework establishes a direct connection between stochastic optimization algorithms and a secondary Bayesian inference problem on gradients, where a prior measure on noisy gradient observations determine the resulting algorithm.

A mail carrier for the United States Postal Service makes deliveries at a Florida apartment complex in June 2018. The USPS has partnered with TuSimple to launch a multi-state driverless semi-truck test program on Tuesday. It doesn't involve home deliveries. A mail carrier for the United States Postal Service makes deliveries at a Florida apartment complex in June 2018. The USPS has partnered with TuSimple to launch a multi-state driverless semi-truck test program on Tuesday.

The Postal Service is experimenting with self-driving long-haul semi trucks to transport mail between distribution centers. The U.S. Postal Service is testing its first long-haul self-driving delivery truck in a two-week pilot program that will use an autonomous tractor-trailer to deliver mail between distribution centers in Phoenix and Dallas. TuSimple, a self-driving truck company, is providing the vehicle and will have a safety engineer and driver in the cab to monitor its performance and take control if there are any issues, the company said in announcing the test Tuesday. The Postal Service has been exploring the idea for some time, recently soliciting bids to put semi-autonomous mail trucks on the road in a few years that allow a human to sort the mail while being autonomously driven along the route. "We are conducting research and testing as part of our efforts to operate a future class of vehicles which will incorporate new technology to accommodate a diverse mail mix, enhance safety, improve service, reduce emissions, and produce operational savings," said Postal Service spokeswoman Kim Frum.

An office worker who believes his image was captured by facial recognition cameras when he popped out for a sandwich in his lunch break has launched a groundbreaking legal battle against the use of the technology. Supported by the campaign group Liberty, Ed Bridges, from Cardiff, raised money through crowdfunding to pursue the action, claiming the suspected use of the technology on him by South Wales police was an unlawful violation of privacy. Bridges, 36, claims he was distressed by the apparent use of the technology and is also arguing during a three-day hearing at Cardiff civil justice and family centre that it breaches data protection and equality laws. Facial recognition technology maps faces in a crowd and then compares them to a watchlist of images, which can include suspects, missing people and persons of interest to the police. The cameras scan faces in large crowds in public places such as streets, shopping centres, football crowds and music events such as the Notting Hill carnival. Bridges, a former Liberal Democrat councillor, believes his image was captured while shopping in Cardiff, and later at a peaceful protest against the arms trade.

A consumer advocacy group has discovered that not all Facebook users have access to a privacy setting that lets them opt out of the site's facial recognition technology. Consumer Reports examined a set of Facebook accounts and found that a significant number didn't have the ability to toggle off Face Recognition, a feature that uses facial recognition technology to identify users in tagged photos. That's despite Facebook announcing almost two years ago that all users would be able to opt out of facial recognition entirely through the setting. A consumer advocacy group has discovered that not all Facebook users have access to a privacy setting that lets them opt out of the site's facial recognition technology Users can control whether they're part of Facebook's facial recognition technology by selecting'privacy shortcuts' in the righthand corner of their News Feed. From there, select'Control face recognition' under Privacy. Select'Edit,' then choose'No' from the dropdown menu.

The group, representing the world's richest countries, hopes non-binding guidelines will become global standard. PARIS -- Donald Trump's administration has finally found an international agreement it can support. At an annual meeting on Wednesday, the 36 countries in the Organization for Economic Cooperation and Development (OECD) plus a handful of other nations are set to adopt a list of guidelines for the development and use of artificial intelligence. The agreement, seen by POLITICO, marks the first time that the United States -- home to some of the world's largest and most powerful tech companies -- has endorsed international guidelines for the emerging technologies. China, the second global front-runner in the field, is not a member of the OECD.

The UK has retained its place among the most prepared governments to harness the opportunities presented by artificial intelligence. An index published today, compiled by Oxford Insights in partnership with the International Development Research Centre (IDRC) in Canada, places the UK as Europe's leading nation and just second on the world stage. "I'm delighted the UK government has been recognised as one of the best in the world in readiness for Artificial Intelligence. AI is already having a positive impact across society – from detecting fraud and diagnosing medical conditions, to helping us discover new music – and we're working hard to make the most of its vast opportunities while managing and mitigating the potential risks. With our newly appointed AI Council, we will boost the growth and use of AI in the UK, by using the knowledge of experts from a range of sectors and encourage dialogue between industry, academia and the public sector, to realise the full potential of data-driven technologies to the economy."

Editor's Note: The following is a guest post from Eric Jensen, head of IoT Product Management at Canonical. In farming, AI is usually short for "artificial insemination." But another kind of AI -- artificial intelligence -- is showing great promise in solving some of agriculture's most significant challenges, from the need to increase productivity and profits to overcoming labor shortages to protecting the environment. Of all the industries AI is transforming, it's safe to say none will have a greater human impact than farming. According to the UN Food and Agriculture Organization, the global population is expected to rise from 7 billion to 9.2 billion by 2050, requiring a 60% increase in food production.