Training deep neural networks for solving machine learning problems is one great challenge in the field, mainly due to its associated optimisation problem being highly non-convex. Recent developments have suggested that many training algorithms do not suffer from undesired local minima under certain scenario, and consequently led to great efforts in pursuing mathematical explanations for such observations. This work provides an alternative mathematical understanding of the challenge from a smooth optimisation perspective. By assuming exact learning of finite samples, sufficient conditions are identified via a critical point analysis to ensure any local minimum to be globally minimal as well. Furthermore, a state of the art algorithm, known as the Generalised Gauss-Newton (GGN) algorithm, is rigorously revisited as an approximate Newton's algorithm, which shares the property of being locally quadratically convergent to a global minimum under the condition of exact learning.

STOCKHOLM/SAN, FRANCISCO – Uber plans to buy up to 24,000 self-driving cars from Volvo, marking the transition of the U.S. firm from an app used to summon a taxi to the owner and operator of a fleet of cars. The nonbinding framework deal could offer San Francisco-based Uber a way to overcome setbacks at its autonomous driving division in Silicon Valley's race to perfect self-driving systems. Combining Volvo's cars with Uber's self-driving system builds on their nearly three-year relationship and comes as Uber's autonomous driving unit has been hit by a lawsuit over trade secrets and the departure of top talent. Automakers, ride-hailing firms and tech start-ups have been forging loose alliances in an effort to advance self-driving technology and claim a piece of what is expected to be a multi-billion-dollar business. Geely-owned Volvo said in a statement on Monday it would provide Uber with its flagship XC90 SUVs equipped with autonomous technology as part of a nonexclusive deal from 2019 to 2021.

For the first time, artificial intelligence (AI) may have eclipsed the science that brought it to "life": we now have AI that can converse in a human-like manner, robots designed to look like us, and deep learning machines built specifically to learn, think, and act the way we do. Experts believe that by 2030, machine intelligence will be on par with humans and that by 2045, the capabilities of AI will actually surpass human intelligence. Given the rate of advancement the field has seen and the resources being dedicated towards its continued development, robotics researchers believe we're closer to "thinking" machines than ever before. "It's getting to a point where we might be able to say this thing has a sense of itself, and maybe there is a threshold moment where suddenly this consciousness emerges," mathematician Marcus du Sautoy from the University of Oxford said. "And if we understand these things are having a level of consciousness, we might well have to introduce rights.

Graphcore has today announced a $50 million Series C funding round by Sequoia Capital as the machine intelligence company prepares to ship its first Intelligence Processing Unit (IPU) products to early access customers at the start of 2018. The Series C round enables Graphcore to significantly accelerate growth to meet the expected global demand for its machine intelligence processor. The funding will be dedicated to scaling up production, building a community of developers around the Poplar software platform, driving Graphcore's extended product roadmap, and investing in its Palo Alto-based US team to help support customers. Nigel Toon, CEO at Graphcore said: "Efficient AI processing power is rapidly becoming the most sought-after resource in the technological world. We believe our IPU technology will become the worldwide standard for machine intelligence compute. The performance of Graphcore's processor, compared to other accelerators, is going to be transformative, whether you are a medical researcher, roboticist, online marketplace, social network or building autonomous vehicles. "At Graphcore, we are focused on building a successful, enduring company that can serve the needs of all of those communities, over the long term.

How can companies identify--and source--the technologies that will be critical for crafting a strategy to keep up in the shifting automotive landscape? The automotive industry is in the early phases of what is expected to be rapid and fundamental change. Technology is the key to further penetration of all these trends, as well as the developing business models that allow companies to capitalize on them. The industry players--traditional automotive companies and new entrants alike--that identify and secure those technological resources will be best positioned to benefit in the new mobility landscape. Thus, industry players need to think about sourcing underlying technologies rather than acquiring single products or services.

People who play multiplayer online battle arenas (MOBA) like Dota 2 and League of Legends perform better on problem solving and logic tests than those who play shooters Destiny and Battlefield 3, researchers found. "The specific MOBA genre is remarkable in the sense that it already engages a vast number of players across the globe, but more generally, complex, socially-interactive and intellectually demanding video games are now ubiquitous and generate a constant stream of performance data that can be normalized against millions of other players," a team from the University of York in the UK concludes. The scientists found that as participants got older, shooter skills dropped. The performance pattern suggested that younger players had an advantage over older ones and that "performance decreases monotonically with age." Since MOBAs tend to favor more strategy and planning than twitch reflexes, this might not be too surprising.

Uber has just taken another big step from a ride-sharing service to a transportation provider. It announced that it will buy up to 24,000 Volvo XC90s, marking the first major vehicle fleet purchase by a ride-hailing service. Uber will take delivery of the SUVs between 2019 and 2021, then equip them with its own sensors and tech, allowing it to do fully autonomous, driver-free passenger rides. "This new agreement puts us on a path toward mass-produced, self-driving vehicles at scale," Uber's Jeff Miller told Bloomberg. The XC90 starts at $47,000, so this could be a pretty substantial purchase -- over $1 billion worth of cars, to be exact. Uber and Volvo previously signed a $300 million pact, and Volvo, based in Sweden but owned by China's Geely Auto, is using the proceeds to develop its own driverless cars.

Sex robot fans can fulfil almost any weird fetish - from a doll with three boobs, to one with elf ears to one with a cat tail - according to an adult performer who visited a US RealDoll manufacturer. Far from being a niche accessory, there are a surprising number of customers prepared to shell out thousands on sexy cyborgs that provide an emotional connection, the source claimed. Some enthusiastic customers already have harems of super realistic cyborgs that can be'anything they want' - with manufacturers saying they are struggling to meet demand. Sex robot fans can fulfil almost any weird fetish - from a doll with three boobs, to one with elf ears to one with a cat tail, according to an adult performer who visited a US RealDoll manufacturer. As part of her visit, Ms SugarCookie also spoke to cyborg Harmony 2.0 who can be programmed with 18 different personality traits, including'shy' and'sexual.' 'My favourite hobby is talking to you... You're so hot Harriet', Harmony said.

About Justin Wang Using AI to Predict the Future with Scry We place too much faith in experts, especially when it come to predicting the future. Experts, of course, failed to predict key world events like the Trump election, Brexit and so on. Justin Wang wants to change that. Scry is an AI powered startup that helps predictors connect with people who want the predictions. Traders can make billions by being right only 51% of the time.

UK Prime Minister, Theresa May today announced a new £20 million (€22.3 million) fund to help public services use UK expertise in innovative technologies like Artificial Intelligence (AI). Investment in the sector is always welcome and who knows, says Jeremy Cowan, it could take Mrs May's mind off Brexit for a while. Earlier, the Prime Minister met leading digital entrepreneurs and innovators from across the UK, as she announced a series of measures to support the "continued growth and success" of the country's tech sector. There was plenty of talk of being world class and of the Government's "enduring commitment" to this vital industry. The PM and Britain's Chancellor of the Exchequer, Philip Hammond announced a wide-ranging package of support which includes: Mrs May said: "It is absolutely right that this dynamic sector, which makes such an immense contribution to our economic life and to our society, has the full backing of Government. Technology is at the heart of our modern ...