In deterministic optimization, line searches are a standard tool ensuring stability and efficiency. Where only stochastic gradients are available, no direct equivalent has so far been formulated, because uncertain gradients do not allow for a strict sequence of decisions collapsing the search space. We construct a probabilistic line search by combining the structure of existing deterministic methods with notions from Bayesian optimization. Our method retains a Gaussian process surrogate of the univariate optimization objective, and uses a probabilistic belief over the Wolfe conditions to monitor the descent. The algorithm has very low computational cost, and no user-controlled parameters. Experiments show that it effectively removes the need to define a learning rate for stochastic gradient descent.

We investigate contextual online learning with nonparametric (Lipschitz) comparison classes under different assumptions on losses and feedback information. For full information feedback and Lipschitz losses, we design the first explicit algorithm achieving the minimax regret rate (up to log factors). In a partial feedback model motivated by second-price auctions, we obtain algorithms for Lipschitz and semi-Lipschitz losses with regret bounds improving on the known bounds for standard bandit feedback. Our analysis combines novel results for contextual second-price auctions with a novel algorithmic approach based on chaining. When the context space is Euclidean, our chaining approach is efficient and delivers an even better regret bound.

There's a lot of noise about how Robo-Advice is going to disrupt the wealth management sector. Indeed, the UK's FCA received over 40 applications last year to use automated wealth solutions of one sort or another. Many Robo-Advisors tout some level of machine intelligence to foster results, but really offer a generic, pre-set portfolio of three to six ETFs to clients. San Francisco-based AI Labs began building its Vise platform n January 2015. Today it uses a mix of machine learning and deep learning.

IBM's Watson can apparently do everything. From manufacturing and medical treatment planning to portrait drawing and filing your taxes, there seems to be no limit to what the Jeopardy-winning AI can do. And next week, Watson will be offering its services to the Wimbledon tennis tournament. Those attending the event will be able to access a Watson-driven digital assistant named Fred via a mobile app. Fred will be able to help them navigate the courts, find food stands and vendors as well as figure out who is playing at any given time.

A logo of Alibaba Group is pictured at its headquarters in Hangzhou, Zhejiang province, China, October 14, 2015. "Amazon Echo", launched in 2014, is a speaker which one can leave on all day and give voice commands to, similar to Siri on an Apple Inc iPhone. Alibaba's new product would be made available only in China and speak only Mandarin, the source told Reuters. Apple and Google, a unit of Alphabet Inc, have unveiled products similar to Echo with the HomePod and Google Home. The Information, a technology website, was the first to report the news on Thursday.

The adoption and application of Artificial Intelligence (AI) continues to accelerate at an exponential rate in modern businesses. As referenced in the 2017 Tech Trend Report, AI is nearing completion of the next layer in technological advancement, integrated into everything individuals and organizations do. This trajectory is predicted to drive cumulative worldwide spending of $40.6 billion on AI projects by 2024 – according to Raconteur. This is expected to create mass opportunity for the pioneering businesses currently investing in AI development. Moving beyond the hype in existing media coverage, this post will uncover the reality behind what AI means for businesses today, in the near future, and beyond 2017.

Predictive analytics functionality is also going to be tested out for the first time. Dimension Data is helping to bring machine learning to the Tour de France. The world famous bike race, which beings in Düsseldorf on Saturday and finishes at the Champs-Élysées in Paris on the 23rd July, is for the first time having machine learning algorithms applied to it. The Dimension Data analytics platform, which was developed in partnership with A.S.O, the organisers of the Tour de France, applies machine learning and algorithms that combine live and historical race data to provide a deeper level of insight into the race. Fans will be able to see rider profiles and learn more about the environments and circumstances in which they perform best.

EARLIER this year Françoise Hardy, a French musician, appeared in a YouTube video (see link). She is asked, by a presenter off-screen, why President Donald Trump sent his press secretary, Sean Spicer, to lie about the size of the inauguration crowd. Then she says Mr Spicer "gave alternative facts to that". It's all a little odd, not least because Françoise Hardy (pictured), who is now 73, looks only 20, and the voice coming out of her mouth belongs to Kellyanne Conway, an adviser to Mr Trump. The video, called "Alternative Face v1.1", is the work of Mario Klingemann, a German artist.

A pizzeria is using super-speedy robots to make fresh pizzas - and for Mountain View residents it could mean dinner is at their door just four minutes after being ordered. California's Zume pizza has pioneered a robot-assisted technique for making the perfect pizza base in just nine seconds. The'Doughbot' is designed to replace the person who spins dough which shaves off 36 seconds of the pizza-making process. Zume pizza has now created a robot-assisted technique for making the perfect pizza base. Silicon Valley's Zume Pizza, which began delivery in April 2016, is using intelligent machines to grab a slice of the multibillion-dollar pizza delivery market.

John Chambers, Chairman of Cisco System 3. Welcome to the Digital Economy Compass 3 Less talking, more facts – our idea behind creating the Digital Economy Compass. It contains facts, trends and key players, covering the entire digital economy. This very first edition will provide everything you need to know about the digital economy. In a global comparison, broadband speed is fastest in East Asia and Scandinavia Source: Akamai Technologies Note: Figure refers to Q3-2016; countries covered: broadband ranking – 147 1: Megabits per second 10 Average broadband speed in Top10 and selected countries (in Mbps1) South Korea 1 Hong Kong 2 Norway 3 23.6 18.2 Sweden 4 20.1 18.4 20.0 19.7 The sevenfold increase in global mobile data traffic is mainly driven by online video streaming Source: Cisco System; figures include only cellular mobile traffic (Wi-Fi or small cell from dual-mode devices are excluded) Note: Other mobile devices include Tablets, mobile PCs, M2M (incl. Pokémon Go was the most popular mobile game for iPhone users across the globe In cooperation with 19 Most downloaded iPhone apps per category and country in 2016 Categories covered: Shopping, Music, News, Gaming, Social Networks U.S. PandoraAmazon CNN Pokémon Go Messenger China Kugou Music Taobao Toutiao WeChat King of Glory Germany SpotifyKleinan zeigen Spiegel Online Pokémon Go WhatsApp U.K. SpotifyeBay BBC News Pokémon Go WhatsApp France DeezerWish Le Monde Pokémon Go Messenger Connectivity 21. "If you make customers unhappy in the physical world, they might each tell 6 friends. If you make customers unhappy on the Internet, they can each tell 6,000 friends."